maximuspowers/muat-fourier-3-classifier
Updated
Datasets:
example_id
int64 0
10k
| metadata
stringlengths 679
723
| classification_prompt
stringlengths 4.51k
10.9k
| classification_completion
stringclasses 14
values | classification_text
stringlengths 4.52k
11k
| improved_signature
stringlengths 2.44k
4.91k
| improved_model_weights
stringlengths 1.76k
5.03k
| training_metrics
stringlengths 1.46k
2.92k
|
|---|---|---|---|---|---|---|---|
0
|
{"target_pattern": "sorted_descending", "degraded_accuracy": 0.4, "improved_accuracy": 0.96, "improvement": 0.5599999999999999, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 5, "neurons_per_layer": 5, "activation_type": "gelu", "dropout_rate": 0.0, "random_seed": 9016, "learning_rate": 0.08961895813761998, "batch_size": 128, "num_epochs": 15, "patience": 3}, "corruption_stats": {"target_pattern": "sorted_descending", "corruption_rate": 0.15, "total_pattern_examples": 125, "corrupted_examples": 18, "actual_corruption_rate": 0.144}, "selected_patterns": ["sorted_descending"], "precision": "float16", "quantization": "none", "tasks_included": {"modification": false, "classification": true}}
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 5
Neurons per Layer: 5
Activation Function: gelu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
-0.135936,
-0.898229,
0.440865,
0.123276,
0.184175
],
[
-0.163952,
-0.131368,
-0.137423,
-0.442275,
0.483762
],
[
0.553204,
0.234059,
0.012346,
-0.007213,
0.122724
],
[
-0.328835,
-0.385777,
0.280825,
0.310734,
0.156388
],
[
-0.605177,
0.107524,
0.407967,
0.51625,
0.077238
]
],
"network.0.bias": [
0.510683,
0.677385,
-0.008098,
-0.082661,
-0.329981
],
"network.2.weight": [
[
0.705301,
0.539599,
-0.01023,
0.021263,
0.746458
],
[
0.717342,
0.678719,
-0.404586,
0.562974,
0.579811
],
[
-0.511617,
0.060623,
-0.159728,
-0.452444,
0.218445
],
[
-0.448086,
-0.305107,
0.722371,
0.410896,
-0.181677
],
[
0.578319,
0.256954,
-0.120081,
0.68794,
0.484082
]
],
"network.2.bias": [
0.131479,
0.382182,
-0.381269,
0.463524,
0.364235
],
"network.4.weight": [
[
-0.831636,
-0.385279,
0.543504,
-0.083977,
-0.008371
],
[
0.159931,
0.352312,
0.416414,
-0.318751,
-0.120372
],
[
0.799975,
0.836729,
-0.107382,
-0.65369,
0.695664
],
[
0.499683,
0.266432,
0.491053,
-0.168315,
0.040631
],
[
0.283262,
0.405537,
0.311068,
-0.301315,
0.387992
]
],
"network.4.bias": [
-0.228648,
-0.175826,
0.188452,
-0.251152,
-0.320186
],
"network.6.weight": [
[
-0.019382,
-0.407566,
-0.092219,
0.112195,
-0.282367
],
[
-0.310099,
-0.194351,
0.106869,
-0.48612,
-0.641673
],
[
-0.029558,
0.361442,
0.717049,
0.360372,
0.199949
],
[
-0.533065,
-0.525414,
-0.063194,
-0.600189,
-0.487264
],
[
-0.389484,
-0.349025,
-0.05247,
0.198462,
-0.528495
]
],
"network.6.bias": [
0.30419,
0.143896,
0.012538,
0.498729,
-0.227905
],
"network.8.weight": [
[
-0.296301,
0.271741,
-0.212339,
0.197531,
-0.171521
],
[
0.175985,
-0.076095,
0.152447,
-0.273649,
0.18578
],
[
-0.510613,
-0.514702,
0.728803,
-0.172438,
0.068894
],
[
-0.16405,
0.226422,
-0.217836,
0.257536,
0.267081
],
[
-0.51732,
-0.459982,
0.77776,
-0.639389,
0.091634
]
],
"network.8.bias": [
-0.271933,
0.027405,
-0.291421,
-0.241338,
-0.102456
],
"network.10.weight": [
[
-0.045999,
0.006641,
-0.610801,
-0.056858,
-0.786437
]
],
"network.10.bias": [
0.158247
]
}
## Activation Signature
### 0
fourier: [[34.383204, 40.171638, 40.350547], [22.239335, 23.422166, 36.439215], [24.708134, 25.798288, 114.167165], [20.733268, 22.589542, 25.321610], [29.322260, 30.772489, 105.063039]]
### 2
fourier: [[26.983542, 30.706222, 156.896161], [35.392268, 36.604519, 145.906124], [12.166634, 14.244054, 74.879210], [18.401049, 19.357862, 81.456438], [30.354496, 31.899994, 150.638379]]
### 4
fourier: [[35.891962, 36.976226, 220.481959], [15.819824, 17.162933, 17.979557], [77.692332, 82.350491, 323.400358], [24.942405, 28.378738, 83.535603], [36.321304, 38.649625, 108.816051]]
### 6
fourier: [[17.765842, 18.558664, 43.137447], [27.580222, 28.680641, 78.077336], [72.635444, 73.128317, 320.971643], [41.956065, 43.070383, 105.774383], [20.534104, 21.408613, 95.924661]]
### 8
fourier: [[15.930071, 16.214262, 87.811183], [11.719948, 12.111654, 45.786713], [54.680187, 55.868834, 201.421832], [16.249771, 17.039159, 89.088282], [59.223098, 61.473636, 227.025592]]
### 10
fourier: [[76.864729, 78.751975, 300.674100]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
|
sorted_descending
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 5
Neurons per Layer: 5
Activation Function: gelu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
-0.135936,
-0.898229,
0.440865,
0.123276,
0.184175
],
[
-0.163952,
-0.131368,
-0.137423,
-0.442275,
0.483762
],
[
0.553204,
0.234059,
0.012346,
-0.007213,
0.122724
],
[
-0.328835,
-0.385777,
0.280825,
0.310734,
0.156388
],
[
-0.605177,
0.107524,
0.407967,
0.51625,
0.077238
]
],
"network.0.bias": [
0.510683,
0.677385,
-0.008098,
-0.082661,
-0.329981
],
"network.2.weight": [
[
0.705301,
0.539599,
-0.01023,
0.021263,
0.746458
],
[
0.717342,
0.678719,
-0.404586,
0.562974,
0.579811
],
[
-0.511617,
0.060623,
-0.159728,
-0.452444,
0.218445
],
[
-0.448086,
-0.305107,
0.722371,
0.410896,
-0.181677
],
[
0.578319,
0.256954,
-0.120081,
0.68794,
0.484082
]
],
"network.2.bias": [
0.131479,
0.382182,
-0.381269,
0.463524,
0.364235
],
"network.4.weight": [
[
-0.831636,
-0.385279,
0.543504,
-0.083977,
-0.008371
],
[
0.159931,
0.352312,
0.416414,
-0.318751,
-0.120372
],
[
0.799975,
0.836729,
-0.107382,
-0.65369,
0.695664
],
[
0.499683,
0.266432,
0.491053,
-0.168315,
0.040631
],
[
0.283262,
0.405537,
0.311068,
-0.301315,
0.387992
]
],
"network.4.bias": [
-0.228648,
-0.175826,
0.188452,
-0.251152,
-0.320186
],
"network.6.weight": [
[
-0.019382,
-0.407566,
-0.092219,
0.112195,
-0.282367
],
[
-0.310099,
-0.194351,
0.106869,
-0.48612,
-0.641673
],
[
-0.029558,
0.361442,
0.717049,
0.360372,
0.199949
],
[
-0.533065,
-0.525414,
-0.063194,
-0.600189,
-0.487264
],
[
-0.389484,
-0.349025,
-0.05247,
0.198462,
-0.528495
]
],
"network.6.bias": [
0.30419,
0.143896,
0.012538,
0.498729,
-0.227905
],
"network.8.weight": [
[
-0.296301,
0.271741,
-0.212339,
0.197531,
-0.171521
],
[
0.175985,
-0.076095,
0.152447,
-0.273649,
0.18578
],
[
-0.510613,
-0.514702,
0.728803,
-0.172438,
0.068894
],
[
-0.16405,
0.226422,
-0.217836,
0.257536,
0.267081
],
[
-0.51732,
-0.459982,
0.77776,
-0.639389,
0.091634
]
],
"network.8.bias": [
-0.271933,
0.027405,
-0.291421,
-0.241338,
-0.102456
],
"network.10.weight": [
[
-0.045999,
0.006641,
-0.610801,
-0.056858,
-0.786437
]
],
"network.10.bias": [
0.158247
]
}
## Activation Signature
### 0
fourier: [[34.383204, 40.171638, 40.350547], [22.239335, 23.422166, 36.439215], [24.708134, 25.798288, 114.167165], [20.733268, 22.589542, 25.321610], [29.322260, 30.772489, 105.063039]]
### 2
fourier: [[26.983542, 30.706222, 156.896161], [35.392268, 36.604519, 145.906124], [12.166634, 14.244054, 74.879210], [18.401049, 19.357862, 81.456438], [30.354496, 31.899994, 150.638379]]
### 4
fourier: [[35.891962, 36.976226, 220.481959], [15.819824, 17.162933, 17.979557], [77.692332, 82.350491, 323.400358], [24.942405, 28.378738, 83.535603], [36.321304, 38.649625, 108.816051]]
### 6
fourier: [[17.765842, 18.558664, 43.137447], [27.580222, 28.680641, 78.077336], [72.635444, 73.128317, 320.971643], [41.956065, 43.070383, 105.774383], [20.534104, 21.408613, 95.924661]]
### 8
fourier: [[15.930071, 16.214262, 87.811183], [11.719948, 12.111654, 45.786713], [54.680187, 55.868834, 201.421832], [16.249771, 17.039159, 89.088282], [59.223098, 61.473636, 227.025592]]
### 10
fourier: [[76.864729, 78.751975, 300.674100]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
sorted_descending
|
{"neuron_activations": {"0": {"neuron_profiles": {"0": {"fourier": [34.38320379607026, 40.171638428931615, 40.35054685021329]}, "1": {"fourier": [22.239334893166323, 23.422165914417118, 36.439215302467346]}, "2": {"fourier": [24.708134034066976, 25.798288262531763, 114.1671646386385]}, "3": {"fourier": [20.733267764821008, 22.589542105793953, 25.321610129533948]}, "4": {"fourier": [29.32226049742522, 30.77248937422029, 105.06303949654102]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "2": {"neuron_profiles": {"0": {"fourier": [26.983542456877764, 30.706222016143478, 156.8961609452963]}, "1": {"fourier": [35.392268467454066, 36.60451876552779, 145.90612418949604]}, "2": {"fourier": [12.16663405960166, 14.244054244718246, 74.87920969724655]}, "3": {"fourier": [18.401048568854097, 19.35786244553081, 81.45643821358681]}, "4": {"fourier": [30.35449610410125, 31.89999351256122, 150.63837936520576]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "4": {"neuron_profiles": {"0": {"fourier": [35.89196153555125, 36.97622649102833, 220.48195853829384]}, "1": {"fourier": [15.819823590633568, 17.16293277608725, 17.979556851089]}, "2": {"fourier": [77.69233215881653, 82.35049057326738, 323.40035815536976]}, "3": {"fourier": [24.9424050654011, 28.37873771312005, 83.53560316562653]}, "4": {"fourier": [36.321303874414944, 38.64962513482666, 108.8160510957241]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "6": {"neuron_profiles": {"0": {"fourier": [17.765841511260682, 18.55866378065493, 43.137447372078896]}, "1": {"fourier": [27.58022228954291, 28.680640880817172, 78.07733555138111]}, "2": {"fourier": [72.63544416145571, 73.12831693209445, 320.9716428425163]}, "3": {"fourier": [41.95606462020902, 43.07038306629363, 105.77438279986382]}, "4": {"fourier": [20.534104444125308, 21.4086128303517, 95.92466147989035]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "8": {"neuron_profiles": {"0": {"fourier": [15.930070809228903, 16.214261714513444, 87.81118267774582]}, "1": {"fourier": [11.719947855432828, 12.111653666811062, 45.78671268187463]}, "2": {"fourier": [54.68018739708468, 55.868833800578486, 201.42183224856853]}, "3": {"fourier": [16.249771396118632, 17.03915850581894, 89.0882818698883]}, "4": {"fourier": [59.22309793909146, 61.47363577980473, 227.02559154108167]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "10": {"neuron_profiles": {"0": {"fourier": [76.8647285612855, 78.75197521340966, 300.6741001754999]}}, "layer_info": {"num_neurons": 1, "num_examples": 90, "profile_methods": ["fourier"]}}}, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 5, "neurons_per_layer": 5, "activation_type": "gelu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}}
|
{"config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 5, "neurons_per_layer": 5, "activation_type": "gelu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}, "weights": {"network.0.weight": [[-0.135936, -0.898229, 0.440865, 0.123276, 0.184175], [-0.163952, -0.131368, -0.137423, -0.442275, 0.483762], [0.553204, 0.234059, 0.012346, -0.007213, 0.122724], [-0.328835, -0.385777, 0.280825, 0.310734, 0.156388], [-0.605177, 0.107524, 0.407967, 0.51625, 0.077238]], "network.0.bias": [0.510683, 0.677385, -0.008098, -0.082661, -0.329981], "network.2.weight": [[0.705301, 0.539599, -0.01023, 0.021263, 0.746458], [0.717342, 0.678719, -0.404586, 0.562974, 0.579811], [-0.511617, 0.060623, -0.159728, -0.452444, 0.218445], [-0.448086, -0.305107, 0.722371, 0.410896, -0.181677], [0.578319, 0.256954, -0.120081, 0.68794, 0.484082]], "network.2.bias": [0.131479, 0.382182, -0.381269, 0.463524, 0.364235], "network.4.weight": [[-0.831636, -0.385279, 0.543504, -0.083977, -0.008371], [0.159931, 0.352312, 0.416414, -0.318751, -0.120372], [0.799975, 0.836729, -0.107382, -0.65369, 0.695664], [0.499683, 0.266432, 0.491053, -0.168315, 0.040631], [0.283262, 0.405537, 0.311068, -0.301315, 0.387992]], "network.4.bias": [-0.228648, -0.175826, 0.188452, -0.251152, -0.320186], "network.6.weight": [[-0.019382, -0.407566, -0.092219, 0.112195, -0.282367], [-0.310099, -0.194351, 0.106869, -0.48612, -0.641673], [-0.029558, 0.361442, 0.717049, 0.360372, 0.199949], [-0.533065, -0.525414, -0.063194, -0.600189, -0.487264], [-0.389484, -0.349025, -0.05247, 0.198462, -0.528495]], "network.6.bias": [0.30419, 0.143896, 0.012538, 0.498729, -0.227905], "network.8.weight": [[-0.296301, 0.271741, -0.212339, 0.197531, -0.171521], [0.175985, -0.076095, 0.152447, -0.273649, 0.18578], [-0.510613, -0.514702, 0.728803, -0.172438, 0.068894], [-0.16405, 0.226422, -0.217836, 0.257536, 0.267081], [-0.51732, -0.459982, 0.77776, -0.639389, 0.091634]], "network.8.bias": [-0.271933, 0.027405, -0.291421, -0.241338, -0.102456], "network.10.weight": [[-0.045999, 0.006641, -0.610801, -0.056858, -0.786437]], "network.10.bias": [0.158247]}}
|
{"training_history": [{"stage": "degraded", "epoch": 0, "global_epoch": 0, "train_loss": 0.6779561638832092, "train_acc": 0.445, "val_loss": 0.6844655871391296, "val_acc": 0.4}, {"stage": "degraded", "epoch": 1, "global_epoch": 1, "train_loss": 0.5416218340396881, "train_acc": 0.59, "val_loss": 0.5125601887702942, "val_acc": 0.4}, {"stage": "improved", "epoch": 0, "global_epoch": 2, "train_loss": 0.39408907294273376, "train_acc": 0.655, "val_loss": 0.35677871108055115, "val_acc": 0.96}, {"stage": "improved", "epoch": 1, "global_epoch": 3, "train_loss": 0.2828209549188614, "train_acc": 0.955, "val_loss": 0.2948799431324005, "val_acc": 0.96}, {"stage": "improved", "epoch": 2, "global_epoch": 4, "train_loss": 0.22725071012973785, "train_acc": 0.955, "val_loss": 0.24082143604755402, "val_acc": 0.96}, {"stage": "improved", "epoch": 3, "global_epoch": 5, "train_loss": 0.19858575612306595, "train_acc": 0.955, "val_loss": 0.20985378324985504, "val_acc": 0.96}, {"stage": "improved", "epoch": 4, "global_epoch": 6, "train_loss": 0.18956772983074188, "train_acc": 0.945, "val_loss": 0.21755105257034302, "val_acc": 0.94}, {"stage": "improved", "epoch": 5, "global_epoch": 7, "train_loss": 0.18157478421926498, "train_acc": 0.94, "val_loss": 0.20880016684532166, "val_acc": 0.94}, {"stage": "improved", "epoch": 6, "global_epoch": 8, "train_loss": 0.16584746539592743, "train_acc": 0.95, "val_loss": 0.204482764005661, "val_acc": 0.94}, {"stage": "improved", "epoch": 7, "global_epoch": 9, "train_loss": 0.16199803352355957, "train_acc": 0.95, "val_loss": 0.20969551801681519, "val_acc": 0.94}, {"stage": "improved", "epoch": 8, "global_epoch": 10, "train_loss": 0.26545940339565277, "train_acc": 0.945, "val_loss": 0.2017844319343567, "val_acc": 0.94}, {"stage": "improved", "epoch": 9, "global_epoch": 11, "train_loss": 0.203144408762455, "train_acc": 0.935, "val_loss": 0.20167064666748047, "val_acc": 0.94}], "summary": {"total_epochs": 12, "degraded_epochs": 2, "improved_epochs": 10, "patterns": ["sorted_descending"], "degraded_stage": {"initial_val_loss": 0.6844655871391296, "final_val_loss": 0.5125601887702942, "initial_val_acc": 0.4, "final_val_acc": 0.4, "best_val_acc": 0.4}, "improved_stage": {"initial_val_loss": 0.35677871108055115, "final_val_loss": 0.20167064666748047, "initial_val_acc": 0.96, "final_val_acc": 0.94, "best_val_acc": 0.96, "best_epoch": 2}, "improvement": 0.5599999999999999, "first_improvement_epoch": 1}}
|
1
|
{"target_pattern": "palindrome", "degraded_accuracy": 0.54, "improved_accuracy": 0.92, "improvement": 0.38, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 6, "neurons_per_layer": 7, "activation_type": "relu", "dropout_rate": 0.0, "random_seed": 2679, "learning_rate": 0.03008896643339405, "batch_size": 128, "num_epochs": 15, "patience": 3}, "corruption_stats": {"target_pattern": "palindrome", "corruption_rate": 0.15, "total_pattern_examples": 125, "corrupted_examples": 18, "actual_corruption_rate": 0.144}, "selected_patterns": ["palindrome"], "precision": "float16", "quantization": "none", "tasks_included": {"modification": false, "classification": true}}
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 6
Neurons per Layer: 7
Activation Function: relu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
0.125469,
-0.22807,
0.229685,
-0.28477,
-0.072223
],
[
0.568673,
0.057374,
0.065451,
-0.042699,
-0.605954
],
[
0.249031,
0.147676,
-0.169346,
0.24949,
0.667312
],
[
-0.473476,
0.076141,
-0.093008,
0.431775,
-0.657419
],
[
0.285524,
0.080429,
-0.115421,
0.182172,
0.817199
],
[
0.710107,
-0.189389,
-0.079083,
0.344341,
0.633212
],
[
-0.054273,
0.00614,
-0.057302,
-0.44436,
-0.220286
]
],
"network.0.bias": [
0.630073,
-0.127634,
0.475448,
0.116184,
0.119195,
-0.001673,
-0.482158
],
"network.2.weight": [
[
-0.087187,
0.164694,
0.272816,
-0.290646,
0.700657,
0.387527,
0.074296
],
[
-0.175392,
0.457536,
0.367264,
-0.352358,
0.32458,
0.405885,
0.180145
],
[
-0.441343,
0.498286,
0.104979,
-0.315393,
0.280045,
0.671565,
-0.284754
],
[
-0.346153,
0.333924,
0.168156,
-0.184189,
0.0652,
0.666748,
-0.174136
],
[
-0.439547,
0.320737,
0.052312,
-0.403155,
0.419374,
0.059859,
0.06173
],
[
-0.423469,
0.197154,
-0.449059,
-0.100629,
-0.429938,
0.082728,
0.169577
],
[
-0.321659,
0.162583,
-0.454295,
0.152977,
-0.27218,
0.194535,
0.182906
]
],
"network.2.bias": [
-0.086853,
0.152959,
0.249638,
0.141462,
0.060801,
-0.03562,
-0.241457
],
"network.4.weight": [
[
0.10578,
0.097875,
0.13872,
0.223178,
-0.14328,
0.325529,
0.026482
],
[
-0.337622,
-0.177456,
0.212671,
-0.053248,
0.166839,
-0.01583,
0.009006
],
[
0.552865,
0.556139,
0.756597,
0.520099,
0.321702,
-0.178331,
0.051394
],
[
-0.297142,
-0.071886,
0.257613,
-0.342738,
0.116841,
0.251029,
-0.236805
],
[
0.363105,
0.008424,
-0.640911,
0.19778,
-0.199553,
-0.127511,
-0.422148
],
[
0.444967,
0.531354,
0.466258,
0.04871,
0.544862,
-0.139811,
0.010258
],
[
-0.201225,
0.286734,
0.031337,
-0.141088,
0.057019,
-0.289465,
-0.06515
]
],
"network.4.bias": [
0.044227,
-0.357857,
-0.138158,
-0.072947,
0.251627,
-0.222342,
0.479936
],
"network.6.weight": [
[
0.314783,
-0.360857,
-0.099691,
0.151684,
0.453793,
-0.028933,
0.596061
],
[
0.078682,
0.200224,
-0.07585,
-0.367313,
-0.461652,
0.117345,
0.218022
],
[
0.314368,
-0.075892,
0.268086,
-0.115635,
-0.433748,
0.588586,
-0.713785
],
[
-0.026679,
0.048687,
-0.017516,
-0.258455,
0.343984,
-0.023492,
0.16277
],
[
-0.026475,
-0.220489,
0.733669,
0.302964,
-0.129162,
0.261333,
-0.39972
],
[
-0.444032,
0.362949,
0.056518,
-0.321557,
-0.393488,
-0.196118,
0.133509
],
[
0.40012,
-0.193117,
0.380514,
-0.366282,
-0.367698,
0.578967,
-0.404105
]
],
"network.6.bias": [
0.254125,
-0.259396,
0.068044,
0.754287,
-0.062867,
-0.054344,
0.059135
],
"network.8.weight": [
[
0.480701,
-0.341995,
-0.032437,
0.552938,
-0.058583,
0.066752,
0.086191
],
[
0.234134,
0.202157,
0.072488,
0.758658,
-0.239738,
-0.073941,
-0.365062
],
[
-0.119979,
0.229057,
0.041364,
-0.056458,
0.456891,
0.141302,
0.661115
],
[
-0.45526,
0.246762,
0.468704,
-0.490493,
0.440737,
-0.035569,
0.250425
],
[
0.012547,
-0.353709,
0.467767,
0.461303,
0.026767,
0.127998,
0.179996
],
[
-0.273539,
0.222246,
-0.065876,
-0.423594,
-0.354655,
-0.369635,
0.282022
],
[
-0.419075,
-0.026043,
0.342473,
-0.738104,
0.623695,
-0.22199,
0.590027
]
],
"network.8.bias": [
0.569178,
0.462256,
-0.107848,
0.199125,
0.399048,
-0.061277,
0.354973
],
"network.10.weight": [
[
0.524196,
0.549181,
0.120931,
-0.367401,
0.396959,
0.027463,
0.08965
],
[
0.553386,
-0.089661,
0.176877,
0.164919,
0.20124,
-0.222673,
-0.266719
],
[
-0.180499,
-0.178738,
0.356455,
0.214917,
0.202369,
0.336856,
-0.307676
],
[
-0.320524,
-0.251655,
0.051755,
0.293225,
-0.301818,
-0.179426,
-0.354398
],
[
0.289126,
0.037438,
-0.037894,
-0.261906,
-0.0301,
-0.10259,
-0.309351
],
[
-0.477042,
-0.58939,
0.609604,
0.60943,
0.187873,
0.154253,
0.597694
],
[
-0.059992,
0.173675,
-0.034314,
-0.019777,
-0.17461,
0.374467,
-0.26644
]
],
"network.10.bias": [
0.43289,
0.165043,
-0.118111,
0.111755,
-0.21986,
0.331246,
-0.252966
],
"network.12.weight": [
[
0.567076,
0.45531,
-0.180927,
0.052572,
0.105219,
-0.457545,
-0.010897
]
],
"network.12.bias": [
0.036335
]
}
## Activation Signature
### 0
fourier: [[14.851885, 15.327329, 15.898825], [22.963743, 24.609837, 27.696848], [30.389728, 33.531173, 185.286600], [32.807046, 37.212072, 39.117923], [36.152803, 39.965163, 160.065606], [39.580665, 45.616853, 170.198236], [18.561104, 19.326078, 169.451684]]
### 2
fourier: [[50.933050, 54.518762, 213.488830], [40.851182, 47.725220, 200.940555], [42.676696, 49.175564, 192.379577], [35.210211, 40.682903, 162.006271], [23.396689, 24.007300, 69.851260], [25.102777, 28.823359, 155.948633], [17.049347, 20.196609, 115.573141]]
### 4
fourier: [[19.512372, 23.046826, 98.468363], [14.483058, 14.902334, 94.586165], [103.851358, 121.791820, 475.080219], [16.478766, 18.233073, 81.628709], [8.240283, 8.501608, 9.434308], [74.428500, 86.453561, 323.536141], [1.513779, 1.618856, 45.209161]]
### 6
fourier: [[7.378812, 8.278544, 29.386429], [3.512162, 3.948875, 9.145049], [78.298544, 91.419672, 318.625046], [4.976697, 5.413555, 60.739369], [95.236097, 111.486453, 406.113464], [16.475660, 19.540613, 84.174327], [90.982567, 106.092227, 391.402461]]
### 8
fourier: [[6.356748, 6.493128, 99.974771], [55.181087, 63.533879, 122.318207], [107.653094, 124.884235, 445.411970], [106.372113, 122.459073, 405.089285], [52.408559, 60.756234, 297.236286], [9.602522, 11.166172, 95.869932], [145.392408, 167.834821, 572.534210]]
### 10
fourier: [[3.643212, 4.552472, 185.224415], [6.058853, 7.186650, 118.839482], [29.546753, 34.344121, 94.362722], [26.827730, 30.631902, 183.399550], [80.014948, 91.000631, 305.071365], [233.389950, 267.093946, 891.038268], [54.407988, 62.475707, 252.889593]]
### 12
fourier: [[107.488760, 121.710483, 273.642032]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
|
palindrome
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 6
Neurons per Layer: 7
Activation Function: relu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
0.125469,
-0.22807,
0.229685,
-0.28477,
-0.072223
],
[
0.568673,
0.057374,
0.065451,
-0.042699,
-0.605954
],
[
0.249031,
0.147676,
-0.169346,
0.24949,
0.667312
],
[
-0.473476,
0.076141,
-0.093008,
0.431775,
-0.657419
],
[
0.285524,
0.080429,
-0.115421,
0.182172,
0.817199
],
[
0.710107,
-0.189389,
-0.079083,
0.344341,
0.633212
],
[
-0.054273,
0.00614,
-0.057302,
-0.44436,
-0.220286
]
],
"network.0.bias": [
0.630073,
-0.127634,
0.475448,
0.116184,
0.119195,
-0.001673,
-0.482158
],
"network.2.weight": [
[
-0.087187,
0.164694,
0.272816,
-0.290646,
0.700657,
0.387527,
0.074296
],
[
-0.175392,
0.457536,
0.367264,
-0.352358,
0.32458,
0.405885,
0.180145
],
[
-0.441343,
0.498286,
0.104979,
-0.315393,
0.280045,
0.671565,
-0.284754
],
[
-0.346153,
0.333924,
0.168156,
-0.184189,
0.0652,
0.666748,
-0.174136
],
[
-0.439547,
0.320737,
0.052312,
-0.403155,
0.419374,
0.059859,
0.06173
],
[
-0.423469,
0.197154,
-0.449059,
-0.100629,
-0.429938,
0.082728,
0.169577
],
[
-0.321659,
0.162583,
-0.454295,
0.152977,
-0.27218,
0.194535,
0.182906
]
],
"network.2.bias": [
-0.086853,
0.152959,
0.249638,
0.141462,
0.060801,
-0.03562,
-0.241457
],
"network.4.weight": [
[
0.10578,
0.097875,
0.13872,
0.223178,
-0.14328,
0.325529,
0.026482
],
[
-0.337622,
-0.177456,
0.212671,
-0.053248,
0.166839,
-0.01583,
0.009006
],
[
0.552865,
0.556139,
0.756597,
0.520099,
0.321702,
-0.178331,
0.051394
],
[
-0.297142,
-0.071886,
0.257613,
-0.342738,
0.116841,
0.251029,
-0.236805
],
[
0.363105,
0.008424,
-0.640911,
0.19778,
-0.199553,
-0.127511,
-0.422148
],
[
0.444967,
0.531354,
0.466258,
0.04871,
0.544862,
-0.139811,
0.010258
],
[
-0.201225,
0.286734,
0.031337,
-0.141088,
0.057019,
-0.289465,
-0.06515
]
],
"network.4.bias": [
0.044227,
-0.357857,
-0.138158,
-0.072947,
0.251627,
-0.222342,
0.479936
],
"network.6.weight": [
[
0.314783,
-0.360857,
-0.099691,
0.151684,
0.453793,
-0.028933,
0.596061
],
[
0.078682,
0.200224,
-0.07585,
-0.367313,
-0.461652,
0.117345,
0.218022
],
[
0.314368,
-0.075892,
0.268086,
-0.115635,
-0.433748,
0.588586,
-0.713785
],
[
-0.026679,
0.048687,
-0.017516,
-0.258455,
0.343984,
-0.023492,
0.16277
],
[
-0.026475,
-0.220489,
0.733669,
0.302964,
-0.129162,
0.261333,
-0.39972
],
[
-0.444032,
0.362949,
0.056518,
-0.321557,
-0.393488,
-0.196118,
0.133509
],
[
0.40012,
-0.193117,
0.380514,
-0.366282,
-0.367698,
0.578967,
-0.404105
]
],
"network.6.bias": [
0.254125,
-0.259396,
0.068044,
0.754287,
-0.062867,
-0.054344,
0.059135
],
"network.8.weight": [
[
0.480701,
-0.341995,
-0.032437,
0.552938,
-0.058583,
0.066752,
0.086191
],
[
0.234134,
0.202157,
0.072488,
0.758658,
-0.239738,
-0.073941,
-0.365062
],
[
-0.119979,
0.229057,
0.041364,
-0.056458,
0.456891,
0.141302,
0.661115
],
[
-0.45526,
0.246762,
0.468704,
-0.490493,
0.440737,
-0.035569,
0.250425
],
[
0.012547,
-0.353709,
0.467767,
0.461303,
0.026767,
0.127998,
0.179996
],
[
-0.273539,
0.222246,
-0.065876,
-0.423594,
-0.354655,
-0.369635,
0.282022
],
[
-0.419075,
-0.026043,
0.342473,
-0.738104,
0.623695,
-0.22199,
0.590027
]
],
"network.8.bias": [
0.569178,
0.462256,
-0.107848,
0.199125,
0.399048,
-0.061277,
0.354973
],
"network.10.weight": [
[
0.524196,
0.549181,
0.120931,
-0.367401,
0.396959,
0.027463,
0.08965
],
[
0.553386,
-0.089661,
0.176877,
0.164919,
0.20124,
-0.222673,
-0.266719
],
[
-0.180499,
-0.178738,
0.356455,
0.214917,
0.202369,
0.336856,
-0.307676
],
[
-0.320524,
-0.251655,
0.051755,
0.293225,
-0.301818,
-0.179426,
-0.354398
],
[
0.289126,
0.037438,
-0.037894,
-0.261906,
-0.0301,
-0.10259,
-0.309351
],
[
-0.477042,
-0.58939,
0.609604,
0.60943,
0.187873,
0.154253,
0.597694
],
[
-0.059992,
0.173675,
-0.034314,
-0.019777,
-0.17461,
0.374467,
-0.26644
]
],
"network.10.bias": [
0.43289,
0.165043,
-0.118111,
0.111755,
-0.21986,
0.331246,
-0.252966
],
"network.12.weight": [
[
0.567076,
0.45531,
-0.180927,
0.052572,
0.105219,
-0.457545,
-0.010897
]
],
"network.12.bias": [
0.036335
]
}
## Activation Signature
### 0
fourier: [[14.851885, 15.327329, 15.898825], [22.963743, 24.609837, 27.696848], [30.389728, 33.531173, 185.286600], [32.807046, 37.212072, 39.117923], [36.152803, 39.965163, 160.065606], [39.580665, 45.616853, 170.198236], [18.561104, 19.326078, 169.451684]]
### 2
fourier: [[50.933050, 54.518762, 213.488830], [40.851182, 47.725220, 200.940555], [42.676696, 49.175564, 192.379577], [35.210211, 40.682903, 162.006271], [23.396689, 24.007300, 69.851260], [25.102777, 28.823359, 155.948633], [17.049347, 20.196609, 115.573141]]
### 4
fourier: [[19.512372, 23.046826, 98.468363], [14.483058, 14.902334, 94.586165], [103.851358, 121.791820, 475.080219], [16.478766, 18.233073, 81.628709], [8.240283, 8.501608, 9.434308], [74.428500, 86.453561, 323.536141], [1.513779, 1.618856, 45.209161]]
### 6
fourier: [[7.378812, 8.278544, 29.386429], [3.512162, 3.948875, 9.145049], [78.298544, 91.419672, 318.625046], [4.976697, 5.413555, 60.739369], [95.236097, 111.486453, 406.113464], [16.475660, 19.540613, 84.174327], [90.982567, 106.092227, 391.402461]]
### 8
fourier: [[6.356748, 6.493128, 99.974771], [55.181087, 63.533879, 122.318207], [107.653094, 124.884235, 445.411970], [106.372113, 122.459073, 405.089285], [52.408559, 60.756234, 297.236286], [9.602522, 11.166172, 95.869932], [145.392408, 167.834821, 572.534210]]
### 10
fourier: [[3.643212, 4.552472, 185.224415], [6.058853, 7.186650, 118.839482], [29.546753, 34.344121, 94.362722], [26.827730, 30.631902, 183.399550], [80.014948, 91.000631, 305.071365], [233.389950, 267.093946, 891.038268], [54.407988, 62.475707, 252.889593]]
### 12
fourier: [[107.488760, 121.710483, 273.642032]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
palindrome
|
{"neuron_activations": {"0": {"neuron_profiles": {"0": {"fourier": [14.851885406020168, 15.327329002609316, 15.898824652225157]}, "1": {"fourier": [22.96374293503703, 24.609837488875996, 27.6968475576902]}, "2": {"fourier": [30.389728259886475, 33.531172914776825, 185.28660035133362]}, "3": {"fourier": [32.807046299186396, 37.21207223087549, 39.11792269988414]}, "4": {"fourier": [36.15280268766395, 39.96516331345573, 160.0656055212021]}, "5": {"fourier": [39.58066489988728, 45.61685277530159, 170.19823560863733]}, "6": {"fourier": [18.561103973846222, 19.326078348610004, 169.45168387889862]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "2": {"neuron_profiles": {"0": {"fourier": [50.93305010006568, 54.51876177817782, 213.488829895854]}, "1": {"fourier": [40.85118164001332, 47.72522021223289, 200.94055522978306]}, "2": {"fourier": [42.67669610346202, 49.17556358954126, 192.37957678735256]}, "3": {"fourier": [35.210210796054625, 40.682903432995786, 162.00627110898495]}, "4": {"fourier": [23.396689267361165, 24.007299533275496, 69.85125966742635]}, "5": {"fourier": [25.102777015942838, 28.82335884070302, 155.94863265752792]}, "6": {"fourier": [17.04934708851315, 20.19660935715838, 115.57314118742943]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "4": {"neuron_profiles": {"0": {"fourier": [19.512371568484436, 23.046826037144072, 98.46836313605309]}, "1": {"fourier": [14.483058042044041, 14.902334081404389, 94.58616533875465]}, "2": {"fourier": [103.85135784504725, 121.79182025104025, 475.08021853864193]}, "3": {"fourier": [16.478766174839627, 18.23307291502258, 81.62870880961418]}, "4": {"fourier": [8.24028301999672, 8.501608008472852, 9.434308263591216]}, "5": {"fourier": [74.42849952457631, 86.45356121108637, 323.5361412167549]}, "6": {"fourier": [1.5137793744554948, 1.6188555911615172, 45.20916086435318]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "6": {"neuron_profiles": {"0": {"fourier": [7.378811936938766, 8.278544002045527, 29.38642869889736]}, "1": {"fourier": [3.5121617993878984, 3.9488748321707154, 9.145049139857292]}, "2": {"fourier": [78.2985440736462, 91.41967180621938, 318.6250456608832]}, "3": {"fourier": [4.976696974831775, 5.413555335928494, 60.739368945360184]}, "4": {"fourier": [95.23609677512123, 111.4864526385402, 406.1134640350938]}, "5": {"fourier": [16.475659938465938, 19.540613062296057, 84.17432655394077]}, "6": {"fourier": [90.98256738136965, 106.09222685777769, 391.4024610668421]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "8": {"neuron_profiles": {"0": {"fourier": [6.356747508550272, 6.493127963923195, 99.9747708439827]}, "1": {"fourier": [55.18108741312582, 63.53387945166109, 122.3182065486908]}, "2": {"fourier": [107.65309378757155, 124.88423531278838, 445.411969602108]}, "3": {"fourier": [106.37211283382929, 122.45907295408236, 405.0892848819494]}, "4": {"fourier": [52.408558533948195, 60.75623392450584, 297.2362861633301]}, "5": {"fourier": [9.602521548213321, 11.166171886940397, 95.86993163824081]}, "6": {"fourier": [145.39240829984527, 167.8348212187903, 572.5342103093863]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "10": {"neuron_profiles": {"0": {"fourier": [3.643212194906667, 4.552471656075601, 185.22441506385803]}, "1": {"fourier": [6.05885317444087, 7.18664997973918, 118.83948218822479]}, "2": {"fourier": [29.54675304636633, 34.344121428194875, 94.36272190511227]}, "3": {"fourier": [26.82772976841645, 30.631902115817855, 183.39955019950867]}, "4": {"fourier": [80.01494778583218, 91.00063090909461, 305.07136453688145]}, "5": {"fourier": [233.3899504994503, 267.09394591696304, 891.0382679700851]}, "6": {"fourier": [54.407987832949935, 62.47570705130021, 252.88959285616875]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "12": {"neuron_profiles": {"0": {"fourier": [107.48875957181731, 121.71048258216115, 273.6420324742794]}}, "layer_info": {"num_neurons": 1, "num_examples": 90, "profile_methods": ["fourier"]}}}, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 6, "neurons_per_layer": 7, "activation_type": "relu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}}
|
{"config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 6, "neurons_per_layer": 7, "activation_type": "relu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}, "weights": {"network.0.weight": [[0.125469, -0.22807, 0.229685, -0.28477, -0.072223], [0.568673, 0.057374, 0.065451, -0.042699, -0.605954], [0.249031, 0.147676, -0.169346, 0.24949, 0.667312], [-0.473476, 0.076141, -0.093008, 0.431775, -0.657419], [0.285524, 0.080429, -0.115421, 0.182172, 0.817199], [0.710107, -0.189389, -0.079083, 0.344341, 0.633212], [-0.054273, 0.00614, -0.057302, -0.44436, -0.220286]], "network.0.bias": [0.630073, -0.127634, 0.475448, 0.116184, 0.119195, -0.001673, -0.482158], "network.2.weight": [[-0.087187, 0.164694, 0.272816, -0.290646, 0.700657, 0.387527, 0.074296], [-0.175392, 0.457536, 0.367264, -0.352358, 0.32458, 0.405885, 0.180145], [-0.441343, 0.498286, 0.104979, -0.315393, 0.280045, 0.671565, -0.284754], [-0.346153, 0.333924, 0.168156, -0.184189, 0.0652, 0.666748, -0.174136], [-0.439547, 0.320737, 0.052312, -0.403155, 0.419374, 0.059859, 0.06173], [-0.423469, 0.197154, -0.449059, -0.100629, -0.429938, 0.082728, 0.169577], [-0.321659, 0.162583, -0.454295, 0.152977, -0.27218, 0.194535, 0.182906]], "network.2.bias": [-0.086853, 0.152959, 0.249638, 0.141462, 0.060801, -0.03562, -0.241457], "network.4.weight": [[0.10578, 0.097875, 0.13872, 0.223178, -0.14328, 0.325529, 0.026482], [-0.337622, -0.177456, 0.212671, -0.053248, 0.166839, -0.01583, 0.009006], [0.552865, 0.556139, 0.756597, 0.520099, 0.321702, -0.178331, 0.051394], [-0.297142, -0.071886, 0.257613, -0.342738, 0.116841, 0.251029, -0.236805], [0.363105, 0.008424, -0.640911, 0.19778, -0.199553, -0.127511, -0.422148], [0.444967, 0.531354, 0.466258, 0.04871, 0.544862, -0.139811, 0.010258], [-0.201225, 0.286734, 0.031337, -0.141088, 0.057019, -0.289465, -0.06515]], "network.4.bias": [0.044227, -0.357857, -0.138158, -0.072947, 0.251627, -0.222342, 0.479936], "network.6.weight": [[0.314783, -0.360857, -0.099691, 0.151684, 0.453793, -0.028933, 0.596061], [0.078682, 0.200224, -0.07585, -0.367313, -0.461652, 0.117345, 0.218022], [0.314368, -0.075892, 0.268086, -0.115635, -0.433748, 0.588586, -0.713785], [-0.026679, 0.048687, -0.017516, -0.258455, 0.343984, -0.023492, 0.16277], [-0.026475, -0.220489, 0.733669, 0.302964, -0.129162, 0.261333, -0.39972], [-0.444032, 0.362949, 0.056518, -0.321557, -0.393488, -0.196118, 0.133509], [0.40012, -0.193117, 0.380514, -0.366282, -0.367698, 0.578967, -0.404105]], "network.6.bias": [0.254125, -0.259396, 0.068044, 0.754287, -0.062867, -0.054344, 0.059135], "network.8.weight": [[0.480701, -0.341995, -0.032437, 0.552938, -0.058583, 0.066752, 0.086191], [0.234134, 0.202157, 0.072488, 0.758658, -0.239738, -0.073941, -0.365062], [-0.119979, 0.229057, 0.041364, -0.056458, 0.456891, 0.141302, 0.661115], [-0.45526, 0.246762, 0.468704, -0.490493, 0.440737, -0.035569, 0.250425], [0.012547, -0.353709, 0.467767, 0.461303, 0.026767, 0.127998, 0.179996], [-0.273539, 0.222246, -0.065876, -0.423594, -0.354655, -0.369635, 0.282022], [-0.419075, -0.026043, 0.342473, -0.738104, 0.623695, -0.22199, 0.590027]], "network.8.bias": [0.569178, 0.462256, -0.107848, 0.199125, 0.399048, -0.061277, 0.354973], "network.10.weight": [[0.524196, 0.549181, 0.120931, -0.367401, 0.396959, 0.027463, 0.08965], [0.553386, -0.089661, 0.176877, 0.164919, 0.20124, -0.222673, -0.266719], [-0.180499, -0.178738, 0.356455, 0.214917, 0.202369, 0.336856, -0.307676], [-0.320524, -0.251655, 0.051755, 0.293225, -0.301818, -0.179426, -0.354398], [0.289126, 0.037438, -0.037894, -0.261906, -0.0301, -0.10259, -0.309351], [-0.477042, -0.58939, 0.609604, 0.60943, 0.187873, 0.154253, 0.597694], [-0.059992, 0.173675, -0.034314, -0.019777, -0.17461, 0.374467, -0.26644]], "network.10.bias": [0.43289, 0.165043, -0.118111, 0.111755, -0.21986, 0.331246, -0.252966], "network.12.weight": [[0.567076, 0.45531, -0.180927, 0.052572, 0.105219, -0.457545, -0.010897]], "network.12.bias": [0.036335]}}
|
{"training_history": [{"stage": "degraded", "epoch": 0, "global_epoch": 0, "train_loss": 0.6893084645271301, "train_acc": 0.565, "val_loss": 0.6865367293357849, "val_acc": 0.54}, {"stage": "degraded", "epoch": 1, "global_epoch": 1, "train_loss": 0.6810666918754578, "train_acc": 0.565, "val_loss": 0.6784811615943909, "val_acc": 0.54}, {"stage": "degraded", "epoch": 2, "global_epoch": 2, "train_loss": 0.6685593724250793, "train_acc": 0.565, "val_loss": 0.6517688035964966, "val_acc": 0.54}, {"stage": "improved", "epoch": 0, "global_epoch": 3, "train_loss": 0.6524258255958557, "train_acc": 0.49, "val_loss": 0.569166898727417, "val_acc": 0.54}, {"stage": "improved", "epoch": 1, "global_epoch": 4, "train_loss": 0.568692147731781, "train_acc": 0.64, "val_loss": 0.46440890431404114, "val_acc": 0.76}, {"stage": "improved", "epoch": 2, "global_epoch": 5, "train_loss": 0.46779032051563263, "train_acc": 0.825, "val_loss": 0.355616956949234, "val_acc": 0.84}, {"stage": "improved", "epoch": 3, "global_epoch": 6, "train_loss": 0.3691777288913727, "train_acc": 0.86, "val_loss": 0.311698317527771, "val_acc": 0.86}, {"stage": "improved", "epoch": 4, "global_epoch": 7, "train_loss": 0.3332885801792145, "train_acc": 0.875, "val_loss": 0.22795723378658295, "val_acc": 0.92}, {"stage": "improved", "epoch": 5, "global_epoch": 8, "train_loss": 0.26788821816444397, "train_acc": 0.89, "val_loss": 0.2353765070438385, "val_acc": 0.9}, {"stage": "improved", "epoch": 6, "global_epoch": 9, "train_loss": 0.3176712542772293, "train_acc": 0.885, "val_loss": 0.32308822870254517, "val_acc": 0.82}, {"stage": "improved", "epoch": 7, "global_epoch": 10, "train_loss": 0.32575756311416626, "train_acc": 0.83, "val_loss": 0.1803550273180008, "val_acc": 0.9}, {"stage": "improved", "epoch": 8, "global_epoch": 11, "train_loss": 0.3036663830280304, "train_acc": 0.885, "val_loss": 0.18534092605113983, "val_acc": 0.92}, {"stage": "improved", "epoch": 9, "global_epoch": 12, "train_loss": 0.2714708149433136, "train_acc": 0.895, "val_loss": 0.21064597368240356, "val_acc": 0.92}], "summary": {"total_epochs": 13, "degraded_epochs": 3, "improved_epochs": 10, "patterns": ["palindrome"], "degraded_stage": {"initial_val_loss": 0.6865367293357849, "final_val_loss": 0.6517688035964966, "initial_val_acc": 0.54, "final_val_acc": 0.54, "best_val_acc": 0.54}, "improved_stage": {"initial_val_loss": 0.569166898727417, "final_val_loss": 0.21064597368240356, "initial_val_acc": 0.54, "final_val_acc": 0.92, "best_val_acc": 0.92, "best_epoch": 7}, "improvement": 0.38, "first_improvement_epoch": 2}}
|
2
|
{"target_pattern": "alternating", "degraded_accuracy": 0.52, "improved_accuracy": 0.86, "improvement": 0.33999999999999997, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 5, "neurons_per_layer": 7, "activation_type": "gelu", "dropout_rate": 0.0, "random_seed": 9451, "learning_rate": 0.07352170370310572, "batch_size": 128, "num_epochs": 15, "patience": 3}, "corruption_stats": {"target_pattern": "alternating", "corruption_rate": 0.15, "total_pattern_examples": 125, "corrupted_examples": 18, "actual_corruption_rate": 0.144}, "selected_patterns": ["alternating"], "precision": "float16", "quantization": "none", "tasks_included": {"modification": false, "classification": true}}
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 5
Neurons per Layer: 7
Activation Function: gelu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
-0.551271,
-0.05203,
0.138717,
0.324818,
0.35961
],
[
-1.061444,
-0.503966,
-0.005292,
-0.151555,
-0.24208
],
[
-0.314833,
-0.241012,
-0.322078,
0.162798,
0.820414
],
[
-0.500186,
0.510924,
-0.962102,
0.49509,
-0.244913
],
[
-0.061225,
0.679702,
0.330693,
0.507457,
0.179238
],
[
0.042317,
-0.201159,
0.826584,
0.116007,
-0.253519
],
[
0.521588,
-0.915429,
0.455052,
-1.224691,
-0.037623
]
],
"network.0.bias": [
0.124305,
-0.034653,
-0.223149,
0.472755,
0.10406,
-0.18332,
-0.387569
],
"network.2.weight": [
[
-0.042601,
0.522327,
-0.489564,
0.70223,
0.103916,
-0.117905,
0.63961
],
[
0.296918,
-0.587314,
0.483767,
-0.600078,
-0.030457,
0.17886,
-0.816548
],
[
0.002259,
0.029952,
-0.176734,
-0.208311,
-0.396168,
0.143401,
-0.011218
],
[
-0.123373,
-0.267522,
0.977841,
-0.917526,
0.520961,
0.737508,
-0.470839
],
[
0.450231,
-0.210405,
-0.285945,
0.663765,
-0.102788,
-0.359461,
0.645087
],
[
0.351589,
-0.364768,
0.935926,
-0.588428,
0.175075,
0.382866,
-0.4638
],
[
0.1997,
0.234478,
-0.296345,
0.711301,
-0.114499,
-0.125785,
0.310501
]
],
"network.2.bias": [
-0.027267,
0.076584,
0.3094,
0.095454,
0.084853,
0.30651,
0.170205
],
"network.4.weight": [
[
-0.078959,
0.554698,
-0.232973,
0.231949,
-0.261364,
0.715643,
-0.111863
],
[
-0.343264,
0.184756,
0.032713,
0.734977,
-0.526448,
0.253429,
-0.657071
],
[
-0.401259,
0.389532,
0.144516,
0.772047,
-0.486383,
0.170517,
-0.384793
],
[
-0.191145,
-0.510685,
-0.251182,
0.609106,
-0.066641,
-0.379438,
-0.320483
],
[
-0.903985,
0.071087,
-0.261565,
-0.051744,
0.546429,
-0.30718,
-0.35031
],
[
-0.787028,
0.394014,
0.041565,
0.184951,
0.151934,
-0.05244,
-0.20611
],
[
-0.210695,
-0.434277,
-0.27275,
-0.335414,
-0.238961,
-0.652354,
0.33002
]
],
"network.4.bias": [
-0.174941,
0.171552,
0.057714,
-0.176249,
-0.277824,
-0.439258,
-0.493793
],
"network.6.weight": [
[
-0.530545,
-0.312282,
0.000207,
-1.129402,
0.21488,
0.090965,
-0.736759
],
[
0.632073,
0.626108,
0.197443,
0.644134,
-0.480709,
0.129942,
0.436443
],
[
-0.067956,
0.000822,
-0.197543,
-0.582207,
0.439925,
0.454907,
-0.145811
],
[
0.173234,
0.799865,
0.160289,
0.710904,
0.018294,
-0.266757,
0.71049
],
[
-0.606671,
-0.471049,
0.259103,
-0.804312,
-0.264609,
-0.373245,
-0.418079
],
[
0.025263,
-0.063031,
-0.286914,
-0.382901,
-0.710405,
-0.652845,
-0.164302
],
[
0.700168,
0.493489,
0.567421,
0.534715,
-0.301783,
-0.148407,
0.607689
]
],
"network.6.bias": [
0.411646,
-0.312268,
0.461555,
-0.154899,
0.378771,
0.050001,
0.016257
],
"network.8.weight": [
[
0.072404,
0.565219,
0.252447,
0.246436,
-0.024323,
0.264462,
-0.048743
],
[
-0.761892,
0.364706,
-0.611505,
0.276479,
-0.425318,
-0.082477,
0.471603
],
[
0.428477,
-0.887917,
0.144858,
-0.458253,
0.854894,
-0.238806,
0.172426
],
[
-0.588839,
0.695912,
-0.338958,
0.776792,
-0.53166,
0.117393,
0.715092
],
[
1.010065,
-0.57299,
0.055056,
-0.346245,
0.270927,
-0.32213,
-0.218867
],
[
0.440949,
-0.528144,
0.102064,
-0.25411,
0.651101,
-0.198051,
-0.149688
],
[
0.034051,
-0.701981,
0.706676,
-0.421987,
-0.046576,
-0.487505,
-0.427655
]
],
"network.8.bias": [
0.064917,
-0.070935,
0.394746,
-0.332074,
0.598626,
0.694958,
0.594553
],
"network.10.weight": [
[
0.15829,
-0.24894,
0.291136,
-0.678436,
0.417538,
0.741722,
0.482813
]
],
"network.10.bias": [
0.723288
]
}
## Activation Signature
### 0
fourier: [[25.244357, 25.272165, 69.597547], [48.968251, 51.870678, 261.766865], [30.022897, 32.905340, 32.935295], [48.427342, 52.840009, 54.680328], [40.043085, 40.993936, 293.984221], [26.338300, 27.985484, 104.897283], [55.837392, 63.117938, 281.585477]]
### 2
fourier: [[18.521589, 19.276363, 50.356526], [19.394233, 20.147705, 22.035089], [17.851729, 19.620269, 91.940210], [40.485911, 43.096359, 198.776936], [19.487518, 19.579427, 24.509010], [29.038945, 30.751904, 138.947095], [16.295723, 16.650091, 26.103504]]
### 4
fourier: [[39.249205, 41.083122, 134.435886], [48.723950, 53.753581, 147.284509], [47.586959, 51.904086, 147.524885], [17.554351, 17.686233, 19.492713], [11.808143, 12.752008, 115.478164], [18.675900, 19.078702, 39.867086], [33.240452, 36.661944, 229.742472]]
### 6
fourier: [[37.675401, 38.083815, 136.105376], [54.960334, 60.429346, 245.158862], [12.745414, 12.970052, 23.714065], [45.479666, 46.168844, 208.873761], [35.447406, 38.049398, 123.179521], [19.115612, 19.433337, 68.582013], [63.694788, 70.311100, 317.901670]]
### 8
fourier: [[37.371955, 39.677794, 189.993346], [64.609793, 71.337862, 277.435309], [59.256032, 63.874632, 216.409304], [118.816062, 128.795860, 527.065263], [61.709469, 67.579969, 219.202833], [50.689166, 55.633455, 159.592305], [84.150578, 90.854217, 350.660369]]
### 10
fourier: [[94.997292, 105.683004, 292.246114]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
|
alternating
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 5
Neurons per Layer: 7
Activation Function: gelu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
-0.551271,
-0.05203,
0.138717,
0.324818,
0.35961
],
[
-1.061444,
-0.503966,
-0.005292,
-0.151555,
-0.24208
],
[
-0.314833,
-0.241012,
-0.322078,
0.162798,
0.820414
],
[
-0.500186,
0.510924,
-0.962102,
0.49509,
-0.244913
],
[
-0.061225,
0.679702,
0.330693,
0.507457,
0.179238
],
[
0.042317,
-0.201159,
0.826584,
0.116007,
-0.253519
],
[
0.521588,
-0.915429,
0.455052,
-1.224691,
-0.037623
]
],
"network.0.bias": [
0.124305,
-0.034653,
-0.223149,
0.472755,
0.10406,
-0.18332,
-0.387569
],
"network.2.weight": [
[
-0.042601,
0.522327,
-0.489564,
0.70223,
0.103916,
-0.117905,
0.63961
],
[
0.296918,
-0.587314,
0.483767,
-0.600078,
-0.030457,
0.17886,
-0.816548
],
[
0.002259,
0.029952,
-0.176734,
-0.208311,
-0.396168,
0.143401,
-0.011218
],
[
-0.123373,
-0.267522,
0.977841,
-0.917526,
0.520961,
0.737508,
-0.470839
],
[
0.450231,
-0.210405,
-0.285945,
0.663765,
-0.102788,
-0.359461,
0.645087
],
[
0.351589,
-0.364768,
0.935926,
-0.588428,
0.175075,
0.382866,
-0.4638
],
[
0.1997,
0.234478,
-0.296345,
0.711301,
-0.114499,
-0.125785,
0.310501
]
],
"network.2.bias": [
-0.027267,
0.076584,
0.3094,
0.095454,
0.084853,
0.30651,
0.170205
],
"network.4.weight": [
[
-0.078959,
0.554698,
-0.232973,
0.231949,
-0.261364,
0.715643,
-0.111863
],
[
-0.343264,
0.184756,
0.032713,
0.734977,
-0.526448,
0.253429,
-0.657071
],
[
-0.401259,
0.389532,
0.144516,
0.772047,
-0.486383,
0.170517,
-0.384793
],
[
-0.191145,
-0.510685,
-0.251182,
0.609106,
-0.066641,
-0.379438,
-0.320483
],
[
-0.903985,
0.071087,
-0.261565,
-0.051744,
0.546429,
-0.30718,
-0.35031
],
[
-0.787028,
0.394014,
0.041565,
0.184951,
0.151934,
-0.05244,
-0.20611
],
[
-0.210695,
-0.434277,
-0.27275,
-0.335414,
-0.238961,
-0.652354,
0.33002
]
],
"network.4.bias": [
-0.174941,
0.171552,
0.057714,
-0.176249,
-0.277824,
-0.439258,
-0.493793
],
"network.6.weight": [
[
-0.530545,
-0.312282,
0.000207,
-1.129402,
0.21488,
0.090965,
-0.736759
],
[
0.632073,
0.626108,
0.197443,
0.644134,
-0.480709,
0.129942,
0.436443
],
[
-0.067956,
0.000822,
-0.197543,
-0.582207,
0.439925,
0.454907,
-0.145811
],
[
0.173234,
0.799865,
0.160289,
0.710904,
0.018294,
-0.266757,
0.71049
],
[
-0.606671,
-0.471049,
0.259103,
-0.804312,
-0.264609,
-0.373245,
-0.418079
],
[
0.025263,
-0.063031,
-0.286914,
-0.382901,
-0.710405,
-0.652845,
-0.164302
],
[
0.700168,
0.493489,
0.567421,
0.534715,
-0.301783,
-0.148407,
0.607689
]
],
"network.6.bias": [
0.411646,
-0.312268,
0.461555,
-0.154899,
0.378771,
0.050001,
0.016257
],
"network.8.weight": [
[
0.072404,
0.565219,
0.252447,
0.246436,
-0.024323,
0.264462,
-0.048743
],
[
-0.761892,
0.364706,
-0.611505,
0.276479,
-0.425318,
-0.082477,
0.471603
],
[
0.428477,
-0.887917,
0.144858,
-0.458253,
0.854894,
-0.238806,
0.172426
],
[
-0.588839,
0.695912,
-0.338958,
0.776792,
-0.53166,
0.117393,
0.715092
],
[
1.010065,
-0.57299,
0.055056,
-0.346245,
0.270927,
-0.32213,
-0.218867
],
[
0.440949,
-0.528144,
0.102064,
-0.25411,
0.651101,
-0.198051,
-0.149688
],
[
0.034051,
-0.701981,
0.706676,
-0.421987,
-0.046576,
-0.487505,
-0.427655
]
],
"network.8.bias": [
0.064917,
-0.070935,
0.394746,
-0.332074,
0.598626,
0.694958,
0.594553
],
"network.10.weight": [
[
0.15829,
-0.24894,
0.291136,
-0.678436,
0.417538,
0.741722,
0.482813
]
],
"network.10.bias": [
0.723288
]
}
## Activation Signature
### 0
fourier: [[25.244357, 25.272165, 69.597547], [48.968251, 51.870678, 261.766865], [30.022897, 32.905340, 32.935295], [48.427342, 52.840009, 54.680328], [40.043085, 40.993936, 293.984221], [26.338300, 27.985484, 104.897283], [55.837392, 63.117938, 281.585477]]
### 2
fourier: [[18.521589, 19.276363, 50.356526], [19.394233, 20.147705, 22.035089], [17.851729, 19.620269, 91.940210], [40.485911, 43.096359, 198.776936], [19.487518, 19.579427, 24.509010], [29.038945, 30.751904, 138.947095], [16.295723, 16.650091, 26.103504]]
### 4
fourier: [[39.249205, 41.083122, 134.435886], [48.723950, 53.753581, 147.284509], [47.586959, 51.904086, 147.524885], [17.554351, 17.686233, 19.492713], [11.808143, 12.752008, 115.478164], [18.675900, 19.078702, 39.867086], [33.240452, 36.661944, 229.742472]]
### 6
fourier: [[37.675401, 38.083815, 136.105376], [54.960334, 60.429346, 245.158862], [12.745414, 12.970052, 23.714065], [45.479666, 46.168844, 208.873761], [35.447406, 38.049398, 123.179521], [19.115612, 19.433337, 68.582013], [63.694788, 70.311100, 317.901670]]
### 8
fourier: [[37.371955, 39.677794, 189.993346], [64.609793, 71.337862, 277.435309], [59.256032, 63.874632, 216.409304], [118.816062, 128.795860, 527.065263], [61.709469, 67.579969, 219.202833], [50.689166, 55.633455, 159.592305], [84.150578, 90.854217, 350.660369]]
### 10
fourier: [[94.997292, 105.683004, 292.246114]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
alternating
|
{"neuron_activations": {"0": {"neuron_profiles": {"0": {"fourier": [25.244357064938395, 25.272165219150512, 69.59754720330238]}, "1": {"fourier": [48.9682509831439, 51.870677516907044, 261.766864720732]}, "2": {"fourier": [30.022897412038205, 32.90533997602122, 32.93529503047466]}, "3": {"fourier": [48.42734209241126, 52.84000857941661, 54.6803275735822]}, "4": {"fourier": [40.04308502649903, 40.9939355408848, 293.98422111570835]}, "5": {"fourier": [26.338300214048274, 27.985483861109625, 104.89728327095509]}, "6": {"fourier": [55.83739192132204, 63.11793834628959, 281.58547699451447]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "2": {"neuron_profiles": {"0": {"fourier": [18.521589227751786, 19.276363074645158, 50.35652565956116]}, "1": {"fourier": [19.394232771600162, 20.147705175842603, 22.035089133860694]}, "2": {"fourier": [17.85172878903702, 19.620268991875896, 91.94020974636078]}, "3": {"fourier": [40.48591137302813, 43.096359059690535, 198.77693597227335]}, "4": {"fourier": [19.487518444980733, 19.57942723824482, 24.50901011377573]}, "5": {"fourier": [29.038945304650156, 30.7519039728328, 138.94709494709969]}, "6": {"fourier": [16.295723249390782, 16.65009063075444, 26.10350412130356]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "4": {"neuron_profiles": {"0": {"fourier": [39.24920528608692, 41.08312155922379, 134.43588572740555]}, "1": {"fourier": [48.7239501935732, 53.75358067925422, 147.28450892865658]}, "2": {"fourier": [47.58695867738096, 51.904086445098585, 147.52488538622856]}, "3": {"fourier": [17.55435077583168, 17.686232917010784, 19.492712711248977]}, "4": {"fourier": [11.808142949401228, 12.752008110896922, 115.47816440463066]}, "5": {"fourier": [18.675899596594636, 19.078701758342795, 39.867086082696915]}, "6": {"fourier": [33.240452141884454, 36.66194435448031, 229.742471575737]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "6": {"neuron_profiles": {"0": {"fourier": [37.675400931784765, 38.08381541315049, 136.10537581145763]}, "1": {"fourier": [54.9603343384359, 60.42934612514012, 245.15886150300503]}, "2": {"fourier": [12.745413918561068, 12.970052294197995, 23.714064866304398]}, "3": {"fourier": [45.479665705615204, 46.16884358356233, 208.8737608641386]}, "4": {"fourier": [35.447406387727355, 38.049397773067554, 123.17952145636082]}, "5": {"fourier": [19.11561169513824, 19.433337109438522, 68.58201307058334]}, "6": {"fourier": [63.69478844102398, 70.31110020487412, 317.9016699641943]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "8": {"neuron_profiles": {"0": {"fourier": [37.3719550139704, 39.67779422746226, 189.99334612116218]}, "1": {"fourier": [64.6097932558409, 71.33786182865774, 277.4353086054325]}, "2": {"fourier": [59.25603187282246, 63.87463232683757, 216.40930369496346]}, "3": {"fourier": [118.816061974796, 128.79586045169472, 527.0652625262737]}, "4": {"fourier": [61.709469176036876, 67.57996891353731, 219.20283323526382]}, "5": {"fourier": [50.689165988308105, 55.63345482922762, 159.5923052430153]}, "6": {"fourier": [84.1505782020917, 90.85421685176198, 350.6603692173958]}}, "layer_info": {"num_neurons": 7, "num_examples": 90, "profile_methods": ["fourier"]}}, "10": {"neuron_profiles": {"0": {"fourier": [94.99729239787806, 105.68300390647457, 292.24611416459084]}}, "layer_info": {"num_neurons": 1, "num_examples": 90, "profile_methods": ["fourier"]}}}, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 5, "neurons_per_layer": 7, "activation_type": "gelu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}}
|
{"config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 5, "neurons_per_layer": 7, "activation_type": "gelu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}, "weights": {"network.0.weight": [[-0.551271, -0.05203, 0.138717, 0.324818, 0.35961], [-1.061444, -0.503966, -0.005292, -0.151555, -0.24208], [-0.314833, -0.241012, -0.322078, 0.162798, 0.820414], [-0.500186, 0.510924, -0.962102, 0.49509, -0.244913], [-0.061225, 0.679702, 0.330693, 0.507457, 0.179238], [0.042317, -0.201159, 0.826584, 0.116007, -0.253519], [0.521588, -0.915429, 0.455052, -1.224691, -0.037623]], "network.0.bias": [0.124305, -0.034653, -0.223149, 0.472755, 0.10406, -0.18332, -0.387569], "network.2.weight": [[-0.042601, 0.522327, -0.489564, 0.70223, 0.103916, -0.117905, 0.63961], [0.296918, -0.587314, 0.483767, -0.600078, -0.030457, 0.17886, -0.816548], [0.002259, 0.029952, -0.176734, -0.208311, -0.396168, 0.143401, -0.011218], [-0.123373, -0.267522, 0.977841, -0.917526, 0.520961, 0.737508, -0.470839], [0.450231, -0.210405, -0.285945, 0.663765, -0.102788, -0.359461, 0.645087], [0.351589, -0.364768, 0.935926, -0.588428, 0.175075, 0.382866, -0.4638], [0.1997, 0.234478, -0.296345, 0.711301, -0.114499, -0.125785, 0.310501]], "network.2.bias": [-0.027267, 0.076584, 0.3094, 0.095454, 0.084853, 0.30651, 0.170205], "network.4.weight": [[-0.078959, 0.554698, -0.232973, 0.231949, -0.261364, 0.715643, -0.111863], [-0.343264, 0.184756, 0.032713, 0.734977, -0.526448, 0.253429, -0.657071], [-0.401259, 0.389532, 0.144516, 0.772047, -0.486383, 0.170517, -0.384793], [-0.191145, -0.510685, -0.251182, 0.609106, -0.066641, -0.379438, -0.320483], [-0.903985, 0.071087, -0.261565, -0.051744, 0.546429, -0.30718, -0.35031], [-0.787028, 0.394014, 0.041565, 0.184951, 0.151934, -0.05244, -0.20611], [-0.210695, -0.434277, -0.27275, -0.335414, -0.238961, -0.652354, 0.33002]], "network.4.bias": [-0.174941, 0.171552, 0.057714, -0.176249, -0.277824, -0.439258, -0.493793], "network.6.weight": [[-0.530545, -0.312282, 0.000207, -1.129402, 0.21488, 0.090965, -0.736759], [0.632073, 0.626108, 0.197443, 0.644134, -0.480709, 0.129942, 0.436443], [-0.067956, 0.000822, -0.197543, -0.582207, 0.439925, 0.454907, -0.145811], [0.173234, 0.799865, 0.160289, 0.710904, 0.018294, -0.266757, 0.71049], [-0.606671, -0.471049, 0.259103, -0.804312, -0.264609, -0.373245, -0.418079], [0.025263, -0.063031, -0.286914, -0.382901, -0.710405, -0.652845, -0.164302], [0.700168, 0.493489, 0.567421, 0.534715, -0.301783, -0.148407, 0.607689]], "network.6.bias": [0.411646, -0.312268, 0.461555, -0.154899, 0.378771, 0.050001, 0.016257], "network.8.weight": [[0.072404, 0.565219, 0.252447, 0.246436, -0.024323, 0.264462, -0.048743], [-0.761892, 0.364706, -0.611505, 0.276479, -0.425318, -0.082477, 0.471603], [0.428477, -0.887917, 0.144858, -0.458253, 0.854894, -0.238806, 0.172426], [-0.588839, 0.695912, -0.338958, 0.776792, -0.53166, 0.117393, 0.715092], [1.010065, -0.57299, 0.055056, -0.346245, 0.270927, -0.32213, -0.218867], [0.440949, -0.528144, 0.102064, -0.25411, 0.651101, -0.198051, -0.149688], [0.034051, -0.701981, 0.706676, -0.421987, -0.046576, -0.487505, -0.427655]], "network.8.bias": [0.064917, -0.070935, 0.394746, -0.332074, 0.598626, 0.694958, 0.594553], "network.10.weight": [[0.15829, -0.24894, 0.291136, -0.678436, 0.417538, 0.741722, 0.482813]], "network.10.bias": [0.723288]}}
|
{"training_history": [{"stage": "degraded", "epoch": 0, "global_epoch": 0, "train_loss": 0.6916628181934357, "train_acc": 0.545, "val_loss": 0.6948902606964111, "val_acc": 0.52}, {"stage": "degraded", "epoch": 1, "global_epoch": 1, "train_loss": 0.656907469034195, "train_acc": 0.56, "val_loss": 0.6221851110458374, "val_acc": 0.52}, {"stage": "improved", "epoch": 0, "global_epoch": 2, "train_loss": 0.5669578015804291, "train_acc": 0.57, "val_loss": 0.46326887607574463, "val_acc": 0.82}, {"stage": "improved", "epoch": 1, "global_epoch": 3, "train_loss": 0.36081814765930176, "train_acc": 0.935, "val_loss": 0.34990695118904114, "val_acc": 0.84}, {"stage": "improved", "epoch": 2, "global_epoch": 4, "train_loss": 0.33851301670074463, "train_acc": 0.865, "val_loss": 0.4313884973526001, "val_acc": 0.82}, {"stage": "improved", "epoch": 3, "global_epoch": 5, "train_loss": 0.2488829791545868, "train_acc": 0.915, "val_loss": 0.5403603315353394, "val_acc": 0.82}, {"stage": "improved", "epoch": 4, "global_epoch": 6, "train_loss": 0.24245692789554596, "train_acc": 0.93, "val_loss": 0.4776495099067688, "val_acc": 0.86}], "summary": {"total_epochs": 7, "degraded_epochs": 2, "improved_epochs": 5, "patterns": ["alternating"], "degraded_stage": {"initial_val_loss": 0.6948902606964111, "final_val_loss": 0.6221851110458374, "initial_val_acc": 0.52, "final_val_acc": 0.52, "best_val_acc": 0.52}, "improved_stage": {"initial_val_loss": 0.46326887607574463, "final_val_loss": 0.4776495099067688, "initial_val_acc": 0.82, "final_val_acc": 0.86, "best_val_acc": 0.86, "best_epoch": 6}, "improvement": 0.33999999999999997, "first_improvement_epoch": 1}}
|
3
|
{"target_pattern": "increasing_pairs", "degraded_accuracy": 0.5, "improved_accuracy": 0.9, "improvement": 0.4, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 6, "activation_type": "gelu", "dropout_rate": 0.0, "random_seed": 7902, "learning_rate": 0.019119242316001303, "batch_size": 128, "num_epochs": 15, "patience": 3}, "corruption_stats": {"target_pattern": "increasing_pairs", "corruption_rate": 0.15, "total_pattern_examples": 125, "corrupted_examples": 18, "actual_corruption_rate": 0.144}, "selected_patterns": ["increasing_pairs"], "precision": "float16", "quantization": "none", "tasks_included": {"modification": false, "classification": true}}
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 4
Neurons per Layer: 6
Activation Function: gelu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
0.308248,
-0.106554,
0.317253,
-0.270284,
-0.076717
],
[
-0.130905,
0.026878,
-0.542554,
-0.487697,
-0.206054
],
[
0.471833,
0.197985,
0.03069,
-0.204486,
-0.204377
],
[
0.484639,
-0.258715,
0.389787,
-0.268101,
-0.179084
],
[
-0.239469,
0.005306,
0.241567,
0.199968,
-0.297911
],
[
-0.124459,
-0.071256,
0.220915,
0.475926,
0.67989
]
],
"network.0.bias": [
0.015848,
0.214772,
0.30621,
-0.191614,
-0.165518,
0.5968
],
"network.2.weight": [
[
0.124599,
0.234848,
-0.195303,
0.261541,
0.003123,
-0.308994
],
[
-0.331412,
-0.163493,
-0.015405,
0.141672,
0.164427,
-0.338402
],
[
-0.152695,
0.130336,
-0.279607,
-0.304313,
-0.423584,
0.210375
],
[
-0.046642,
-0.208845,
-0.350994,
0.063696,
-0.344599,
0.105945
],
[
-0.061839,
0.247712,
-0.11397,
0.288401,
0.019977,
-0.297684
],
[
0.247831,
0.008517,
-0.154214,
-0.222045,
0.341944,
-0.378961
]
],
"network.2.bias": [
-0.156689,
-0.139741,
-0.119344,
-0.130299,
-0.273499,
0.151582
],
"network.4.weight": [
[
-0.538956,
-0.301866,
-0.561536,
-0.02514,
-0.404367,
-0.497798
],
[
0.105127,
0.135962,
0.193674,
0.376524,
0.19196,
0.610889
],
[
-0.594477,
-0.487278,
-0.197287,
0.116766,
-0.300355,
-0.103628
],
[
-0.399337,
0.121736,
-0.379296,
-0.21849,
-0.182644,
-0.129434
],
[
-0.015705,
0.070121,
0.348506,
-0.188676,
0.138639,
0.537886
],
[
-0.367265,
-0.227509,
-0.19863,
-0.593053,
-0.541858,
-0.457003
]
],
"network.4.bias": [
0.531708,
-0.158666,
-0.073632,
0.456872,
-0.230031,
0.469603
],
"network.6.weight": [
[
0.329071,
-0.498207,
0.332671,
0.230136,
-0.532708,
0.643445
],
[
-0.435571,
0.395488,
-0.308415,
0.195936,
0.154233,
-0.364883
],
[
-0.366484,
0.229135,
-0.296155,
-0.021536,
0.117018,
0.027221
],
[
0.616569,
0.111167,
-0.118406,
0.187173,
0.158521,
0.615021
],
[
0.078434,
0.394491,
-0.053647,
0.212129,
0.345927,
-0.016275
],
[
0.320474,
-0.345668,
0.0394,
0.497217,
-0.655889,
0.190344
]
],
"network.6.bias": [
0.245959,
0.054373,
-0.026117,
0.228214,
-0.064169,
0.226606
],
"network.8.weight": [
[
-0.526297,
0.226127,
0.398418,
-0.247134,
0.001141,
-0.626907
]
],
"network.8.bias": [
0.255489
]
}
## Activation Signature
### 0
fourier: [[19.862496, 20.830531, 23.455965], [29.712725, 31.584885, 209.921728], [19.809704, 24.477208, 56.491790], [24.325519, 28.552861, 31.141662], [14.279831, 14.689653, 17.131972], [31.136291, 34.110272, 236.940058]]
### 2
fourier: [[9.745191, 9.929557, 82.387070], [12.619947, 13.233478, 94.839780], [13.122304, 14.454589, 15.650824], [6.903355, 7.369968, 16.468549], [9.535084, 9.806261, 91.145412], [14.965161, 15.037413, 74.762907]]
### 4
fourier: [[5.158812, 5.856387, 60.326506], [3.188784, 3.624030, 23.633399], [1.688623, 2.290111, 7.728030], [4.073637, 4.140504, 43.246370], [2.257040, 2.810492, 23.866414], [4.303764, 4.712712, 58.447111]]
### 6
fourier: [[5.606738, 6.359093, 83.880835], [3.324797, 3.968807, 31.967752], [2.081108, 2.471892, 23.352558], [4.720023, 5.318862, 79.813636], [0.233760, 0.261278, 2.982098], [4.487900, 4.923892, 68.209054]]
### 8
fourier: [[6.836366, 7.788530, 70.304173]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
|
increasing_pairs
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 4
Neurons per Layer: 6
Activation Function: gelu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
0.308248,
-0.106554,
0.317253,
-0.270284,
-0.076717
],
[
-0.130905,
0.026878,
-0.542554,
-0.487697,
-0.206054
],
[
0.471833,
0.197985,
0.03069,
-0.204486,
-0.204377
],
[
0.484639,
-0.258715,
0.389787,
-0.268101,
-0.179084
],
[
-0.239469,
0.005306,
0.241567,
0.199968,
-0.297911
],
[
-0.124459,
-0.071256,
0.220915,
0.475926,
0.67989
]
],
"network.0.bias": [
0.015848,
0.214772,
0.30621,
-0.191614,
-0.165518,
0.5968
],
"network.2.weight": [
[
0.124599,
0.234848,
-0.195303,
0.261541,
0.003123,
-0.308994
],
[
-0.331412,
-0.163493,
-0.015405,
0.141672,
0.164427,
-0.338402
],
[
-0.152695,
0.130336,
-0.279607,
-0.304313,
-0.423584,
0.210375
],
[
-0.046642,
-0.208845,
-0.350994,
0.063696,
-0.344599,
0.105945
],
[
-0.061839,
0.247712,
-0.11397,
0.288401,
0.019977,
-0.297684
],
[
0.247831,
0.008517,
-0.154214,
-0.222045,
0.341944,
-0.378961
]
],
"network.2.bias": [
-0.156689,
-0.139741,
-0.119344,
-0.130299,
-0.273499,
0.151582
],
"network.4.weight": [
[
-0.538956,
-0.301866,
-0.561536,
-0.02514,
-0.404367,
-0.497798
],
[
0.105127,
0.135962,
0.193674,
0.376524,
0.19196,
0.610889
],
[
-0.594477,
-0.487278,
-0.197287,
0.116766,
-0.300355,
-0.103628
],
[
-0.399337,
0.121736,
-0.379296,
-0.21849,
-0.182644,
-0.129434
],
[
-0.015705,
0.070121,
0.348506,
-0.188676,
0.138639,
0.537886
],
[
-0.367265,
-0.227509,
-0.19863,
-0.593053,
-0.541858,
-0.457003
]
],
"network.4.bias": [
0.531708,
-0.158666,
-0.073632,
0.456872,
-0.230031,
0.469603
],
"network.6.weight": [
[
0.329071,
-0.498207,
0.332671,
0.230136,
-0.532708,
0.643445
],
[
-0.435571,
0.395488,
-0.308415,
0.195936,
0.154233,
-0.364883
],
[
-0.366484,
0.229135,
-0.296155,
-0.021536,
0.117018,
0.027221
],
[
0.616569,
0.111167,
-0.118406,
0.187173,
0.158521,
0.615021
],
[
0.078434,
0.394491,
-0.053647,
0.212129,
0.345927,
-0.016275
],
[
0.320474,
-0.345668,
0.0394,
0.497217,
-0.655889,
0.190344
]
],
"network.6.bias": [
0.245959,
0.054373,
-0.026117,
0.228214,
-0.064169,
0.226606
],
"network.8.weight": [
[
-0.526297,
0.226127,
0.398418,
-0.247134,
0.001141,
-0.626907
]
],
"network.8.bias": [
0.255489
]
}
## Activation Signature
### 0
fourier: [[19.862496, 20.830531, 23.455965], [29.712725, 31.584885, 209.921728], [19.809704, 24.477208, 56.491790], [24.325519, 28.552861, 31.141662], [14.279831, 14.689653, 17.131972], [31.136291, 34.110272, 236.940058]]
### 2
fourier: [[9.745191, 9.929557, 82.387070], [12.619947, 13.233478, 94.839780], [13.122304, 14.454589, 15.650824], [6.903355, 7.369968, 16.468549], [9.535084, 9.806261, 91.145412], [14.965161, 15.037413, 74.762907]]
### 4
fourier: [[5.158812, 5.856387, 60.326506], [3.188784, 3.624030, 23.633399], [1.688623, 2.290111, 7.728030], [4.073637, 4.140504, 43.246370], [2.257040, 2.810492, 23.866414], [4.303764, 4.712712, 58.447111]]
### 6
fourier: [[5.606738, 6.359093, 83.880835], [3.324797, 3.968807, 31.967752], [2.081108, 2.471892, 23.352558], [4.720023, 5.318862, 79.813636], [0.233760, 0.261278, 2.982098], [4.487900, 4.923892, 68.209054]]
### 8
fourier: [[6.836366, 7.788530, 70.304173]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
increasing_pairs
|
{"neuron_activations": {"0": {"neuron_profiles": {"0": {"fourier": [19.862496341990465, 20.830531205643247, 23.455964750442998]}, "1": {"fourier": [29.712725321245426, 31.584885187962293, 209.921728387475]}, "2": {"fourier": [19.809703701121816, 24.477207544049705, 56.49178962409496]}, "3": {"fourier": [24.32551949189835, 28.552861332166607, 31.14166227207478]}, "4": {"fourier": [14.279830885156663, 14.68965345871899, 17.1319721534859]}, "5": {"fourier": [31.136290711930258, 34.11027155000762, 236.94005820155144]}}, "layer_info": {"num_neurons": 6, "num_examples": 90, "profile_methods": ["fourier"]}}, "2": {"neuron_profiles": {"0": {"fourier": [9.745190527610298, 9.929556625790516, 82.38706992566586]}, "1": {"fourier": [12.61994737826265, 13.233477836810962, 94.83977988362312]}, "2": {"fourier": [13.122304276839248, 14.454588573889383, 15.650823926708854]}, "3": {"fourier": [6.903354595228121, 7.369967934524945, 16.468548573553562]}, "4": {"fourier": [9.535083748666587, 9.806261003419548, 91.14541153609753]}, "5": {"fourier": [14.965161305190763, 15.037413277237276, 74.76290714740753]}}, "layer_info": {"num_neurons": 6, "num_examples": 90, "profile_methods": ["fourier"]}}, "4": {"neuron_profiles": {"0": {"fourier": [5.158811801971683, 5.856386962219855, 60.32650563120842]}, "1": {"fourier": [3.1887840447338505, 3.624029569712434, 23.633399114012718]}, "2": {"fourier": [1.6886234188619673, 2.2901106814488355, 7.728029906749725]}, "3": {"fourier": [4.073637328838318, 4.1405041450608175, 43.246370285749435]}, "4": {"fourier": [2.257039626267557, 2.810491875959659, 23.86641366034746]}, "5": {"fourier": [4.303764085341598, 4.712711633767883, 58.447111159563065]}}, "layer_info": {"num_neurons": 6, "num_examples": 90, "profile_methods": ["fourier"]}}, "6": {"neuron_profiles": {"0": {"fourier": [5.606738163454502, 6.359093270389631, 83.88083507120609]}, "1": {"fourier": [3.324797332654858, 3.968807305387114, 31.967751752585173]}, "2": {"fourier": [2.0811077691742383, 2.471892257453908, 23.35255754739046]}, "3": {"fourier": [4.720022961085628, 5.318861896205335, 79.81363618373871]}, "4": {"fourier": [0.23375989508949832, 0.2612782926839001, 2.98209759965539]}, "5": {"fourier": [4.487900273981504, 4.923891626604455, 68.20905402302742]}}, "layer_info": {"num_neurons": 6, "num_examples": 90, "profile_methods": ["fourier"]}}, "8": {"neuron_profiles": {"0": {"fourier": [6.836365559188387, 7.788529974472033, 70.30417285859585]}}, "layer_info": {"num_neurons": 1, "num_examples": 90, "profile_methods": ["fourier"]}}}, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 6, "activation_type": "gelu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}}
|
{"config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 6, "activation_type": "gelu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}, "weights": {"network.0.weight": [[0.308248, -0.106554, 0.317253, -0.270284, -0.076717], [-0.130905, 0.026878, -0.542554, -0.487697, -0.206054], [0.471833, 0.197985, 0.03069, -0.204486, -0.204377], [0.484639, -0.258715, 0.389787, -0.268101, -0.179084], [-0.239469, 0.005306, 0.241567, 0.199968, -0.297911], [-0.124459, -0.071256, 0.220915, 0.475926, 0.67989]], "network.0.bias": [0.015848, 0.214772, 0.30621, -0.191614, -0.165518, 0.5968], "network.2.weight": [[0.124599, 0.234848, -0.195303, 0.261541, 0.003123, -0.308994], [-0.331412, -0.163493, -0.015405, 0.141672, 0.164427, -0.338402], [-0.152695, 0.130336, -0.279607, -0.304313, -0.423584, 0.210375], [-0.046642, -0.208845, -0.350994, 0.063696, -0.344599, 0.105945], [-0.061839, 0.247712, -0.11397, 0.288401, 0.019977, -0.297684], [0.247831, 0.008517, -0.154214, -0.222045, 0.341944, -0.378961]], "network.2.bias": [-0.156689, -0.139741, -0.119344, -0.130299, -0.273499, 0.151582], "network.4.weight": [[-0.538956, -0.301866, -0.561536, -0.02514, -0.404367, -0.497798], [0.105127, 0.135962, 0.193674, 0.376524, 0.19196, 0.610889], [-0.594477, -0.487278, -0.197287, 0.116766, -0.300355, -0.103628], [-0.399337, 0.121736, -0.379296, -0.21849, -0.182644, -0.129434], [-0.015705, 0.070121, 0.348506, -0.188676, 0.138639, 0.537886], [-0.367265, -0.227509, -0.19863, -0.593053, -0.541858, -0.457003]], "network.4.bias": [0.531708, -0.158666, -0.073632, 0.456872, -0.230031, 0.469603], "network.6.weight": [[0.329071, -0.498207, 0.332671, 0.230136, -0.532708, 0.643445], [-0.435571, 0.395488, -0.308415, 0.195936, 0.154233, -0.364883], [-0.366484, 0.229135, -0.296155, -0.021536, 0.117018, 0.027221], [0.616569, 0.111167, -0.118406, 0.187173, 0.158521, 0.615021], [0.078434, 0.394491, -0.053647, 0.212129, 0.345927, -0.016275], [0.320474, -0.345668, 0.0394, 0.497217, -0.655889, 0.190344]], "network.6.bias": [0.245959, 0.054373, -0.026117, 0.228214, -0.064169, 0.226606], "network.8.weight": [[-0.526297, 0.226127, 0.398418, -0.247134, 0.001141, -0.626907]], "network.8.bias": [0.255489]}}
|
{"training_history": [{"stage": "degraded", "epoch": 0, "global_epoch": 0, "train_loss": 0.7180013060569763, "train_acc": 0.425, "val_loss": 0.6934688091278076, "val_acc": 0.5}, {"stage": "degraded", "epoch": 1, "global_epoch": 1, "train_loss": 0.7019679844379425, "train_acc": 0.425, "val_loss": 0.6853039264678955, "val_acc": 0.68}, {"stage": "degraded", "epoch": 2, "global_epoch": 2, "train_loss": 0.6843142807483673, "train_acc": 0.575, "val_loss": 0.6770879626274109, "val_acc": 0.5}, {"stage": "degraded", "epoch": 3, "global_epoch": 3, "train_loss": 0.6714287102222443, "train_acc": 0.575, "val_loss": 0.6651756763458252, "val_acc": 0.5}, {"stage": "degraded", "epoch": 4, "global_epoch": 4, "train_loss": 0.6483409404754639, "train_acc": 0.575, "val_loss": 0.6376786828041077, "val_acc": 0.5}, {"stage": "improved", "epoch": 0, "global_epoch": 5, "train_loss": 0.6328597962856293, "train_acc": 0.53, "val_loss": 0.5665894746780396, "val_acc": 0.84}, {"stage": "improved", "epoch": 1, "global_epoch": 6, "train_loss": 0.572889506816864, "train_acc": 0.775, "val_loss": 0.49300551414489746, "val_acc": 0.9}, {"stage": "improved", "epoch": 2, "global_epoch": 7, "train_loss": 0.5088239163160324, "train_acc": 0.835, "val_loss": 0.4252799153327942, "val_acc": 0.88}, {"stage": "improved", "epoch": 3, "global_epoch": 8, "train_loss": 0.4628315716981888, "train_acc": 0.83, "val_loss": 0.3757992684841156, "val_acc": 0.9}, {"stage": "improved", "epoch": 4, "global_epoch": 9, "train_loss": 0.37355081737041473, "train_acc": 0.85, "val_loss": 0.38032832741737366, "val_acc": 0.84}, {"stage": "improved", "epoch": 5, "global_epoch": 10, "train_loss": 0.36860182881355286, "train_acc": 0.85, "val_loss": 0.3454854488372803, "val_acc": 0.88}, {"stage": "improved", "epoch": 6, "global_epoch": 11, "train_loss": 0.3211270272731781, "train_acc": 0.86, "val_loss": 0.3219808340072632, "val_acc": 0.88}, {"stage": "improved", "epoch": 7, "global_epoch": 12, "train_loss": 0.3318009674549103, "train_acc": 0.85, "val_loss": 0.34785231947898865, "val_acc": 0.88}, {"stage": "improved", "epoch": 8, "global_epoch": 13, "train_loss": 0.35081204771995544, "train_acc": 0.865, "val_loss": 0.3642524778842926, "val_acc": 0.88}, {"stage": "improved", "epoch": 9, "global_epoch": 14, "train_loss": 0.3323000818490982, "train_acc": 0.87, "val_loss": 0.32225313782691956, "val_acc": 0.86}], "summary": {"total_epochs": 15, "degraded_epochs": 5, "improved_epochs": 10, "patterns": ["increasing_pairs"], "degraded_stage": {"initial_val_loss": 0.6934688091278076, "final_val_loss": 0.6376786828041077, "initial_val_acc": 0.5, "final_val_acc": 0.5, "best_val_acc": 0.5}, "improved_stage": {"initial_val_loss": 0.5665894746780396, "final_val_loss": 0.32225313782691956, "initial_val_acc": 0.84, "final_val_acc": 0.86, "best_val_acc": 0.9, "best_epoch": 6}, "improvement": 0.4, "first_improvement_epoch": 4}}
|
4
|
{"target_pattern": "increasing_pairs", "degraded_accuracy": 0.64, "improved_accuracy": 0.86, "improvement": 0.21999999999999997, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 5, "activation_type": "relu", "dropout_rate": 0.0, "random_seed": 9859, "learning_rate": 0.015384002471586396, "batch_size": 128, "num_epochs": 15, "patience": 3}, "corruption_stats": {"target_pattern": "increasing_pairs", "corruption_rate": 0.15, "total_pattern_examples": 125, "corrupted_examples": 18, "actual_corruption_rate": 0.144}, "selected_patterns": ["increasing_pairs"], "precision": "float16", "quantization": "none", "tasks_included": {"modification": false, "classification": true}}
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 4
Neurons per Layer: 5
Activation Function: relu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
0.490115,
-0.014935,
0.309498,
0.129559,
-0.346239
],
[
-0.099157,
-0.083075,
-0.152816,
-0.473807,
-0.371314
],
[
0.770528,
0.127912,
-0.002052,
0.235109,
-0.385436
],
[
0.596325,
0.068987,
0.011688,
-0.296899,
0.164491
],
[
-0.022468,
0.03651,
0.708453,
-0.52115,
-0.20134
]
],
"network.0.bias": [
-0.119875,
0.082194,
0.65578,
0.156972,
-0.069177
],
"network.2.weight": [
[
0.459525,
-0.193837,
0.608708,
0.285532,
0.178922
],
[
-0.121903,
0.002356,
-0.0295,
-0.003489,
0.571354
],
[
0.495116,
-0.121997,
0.556749,
0.148564,
0.403205
],
[
0.743419,
0.077521,
-0.143338,
0.07489,
0.61667
],
[
0.313957,
-0.130418,
-0.15065,
0.341614,
0.108904
]
],
"network.2.bias": [
-0.028663,
0.032932,
-0.072576,
-0.052439,
-0.330283
],
"network.4.weight": [
[
0.028013,
0.312583,
-0.064444,
-0.171507,
0.186128
],
[
0.180827,
0.054785,
-0.261857,
-0.272733,
0.337052
],
[
0.794152,
0.556423,
0.137221,
0.656312,
0.059673
],
[
0.505367,
0.372168,
0.625924,
-0.212468,
0.332403
],
[
0.56674,
-0.081706,
0.542264,
0.157652,
0.524919
]
],
"network.4.bias": [
-0.351487,
-0.300211,
-0.00327,
-0.065848,
0.007807
],
"network.6.weight": [
[
0.390837,
-0.374882,
-0.283566,
-0.381474,
-0.334873
],
[
-0.390806,
-0.397559,
-0.360419,
0.086452,
0.105742
],
[
0.405974,
0.299639,
0.537839,
0.392613,
0.718797
],
[
-0.213986,
-0.03172,
-0.003601,
-0.690389,
0.184278
],
[
0.219681,
0.139951,
-0.188363,
0.643107,
0.629333
]
],
"network.6.bias": [
-0.026605,
-0.14395,
-0.111508,
0.401064,
0.119937
],
"network.8.weight": [
[
-0.366589,
0.108788,
-0.629808,
0.333334,
-0.126564
]
],
"network.8.bias": [
0.144534
]
}
## Activation Signature
### 0
fourier: [[25.760290, 26.184900, 86.412873], [24.915543, 26.225291, 178.722497], [29.024818, 32.599501, 168.503991], [26.189072, 31.088983, 55.384145], [28.173956, 31.290883, 33.072054]]
### 2
fourier: [[33.132748, 37.249733, 180.373191], [11.319644, 11.371465, 20.990937], [31.176552, 36.537869, 174.249795], [22.413155, 22.765480, 87.454885], [10.533273, 11.753672, 12.083696]]
### 4
fourier: [[2.233928, 2.459772, 39.998685], [5.398062, 6.225882, 54.781467], [46.001567, 51.758616, 243.401373], [36.525696, 41.272591, 193.938742], [43.110570, 47.962721, 222.843906]]
### 6
fourier: [[41.198592, 46.410542, 220.087933], [9.149944, 10.104752, 60.336608], [69.716113, 78.398193, 357.265496], [17.362365, 19.808017, 57.728613], [42.092502, 47.005391, 230.024288]]
### 8
fourier: [[49.617480, 55.886713, 239.003486]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
|
increasing_pairs
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 4
Neurons per Layer: 5
Activation Function: relu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
0.490115,
-0.014935,
0.309498,
0.129559,
-0.346239
],
[
-0.099157,
-0.083075,
-0.152816,
-0.473807,
-0.371314
],
[
0.770528,
0.127912,
-0.002052,
0.235109,
-0.385436
],
[
0.596325,
0.068987,
0.011688,
-0.296899,
0.164491
],
[
-0.022468,
0.03651,
0.708453,
-0.52115,
-0.20134
]
],
"network.0.bias": [
-0.119875,
0.082194,
0.65578,
0.156972,
-0.069177
],
"network.2.weight": [
[
0.459525,
-0.193837,
0.608708,
0.285532,
0.178922
],
[
-0.121903,
0.002356,
-0.0295,
-0.003489,
0.571354
],
[
0.495116,
-0.121997,
0.556749,
0.148564,
0.403205
],
[
0.743419,
0.077521,
-0.143338,
0.07489,
0.61667
],
[
0.313957,
-0.130418,
-0.15065,
0.341614,
0.108904
]
],
"network.2.bias": [
-0.028663,
0.032932,
-0.072576,
-0.052439,
-0.330283
],
"network.4.weight": [
[
0.028013,
0.312583,
-0.064444,
-0.171507,
0.186128
],
[
0.180827,
0.054785,
-0.261857,
-0.272733,
0.337052
],
[
0.794152,
0.556423,
0.137221,
0.656312,
0.059673
],
[
0.505367,
0.372168,
0.625924,
-0.212468,
0.332403
],
[
0.56674,
-0.081706,
0.542264,
0.157652,
0.524919
]
],
"network.4.bias": [
-0.351487,
-0.300211,
-0.00327,
-0.065848,
0.007807
],
"network.6.weight": [
[
0.390837,
-0.374882,
-0.283566,
-0.381474,
-0.334873
],
[
-0.390806,
-0.397559,
-0.360419,
0.086452,
0.105742
],
[
0.405974,
0.299639,
0.537839,
0.392613,
0.718797
],
[
-0.213986,
-0.03172,
-0.003601,
-0.690389,
0.184278
],
[
0.219681,
0.139951,
-0.188363,
0.643107,
0.629333
]
],
"network.6.bias": [
-0.026605,
-0.14395,
-0.111508,
0.401064,
0.119937
],
"network.8.weight": [
[
-0.366589,
0.108788,
-0.629808,
0.333334,
-0.126564
]
],
"network.8.bias": [
0.144534
]
}
## Activation Signature
### 0
fourier: [[25.760290, 26.184900, 86.412873], [24.915543, 26.225291, 178.722497], [29.024818, 32.599501, 168.503991], [26.189072, 31.088983, 55.384145], [28.173956, 31.290883, 33.072054]]
### 2
fourier: [[33.132748, 37.249733, 180.373191], [11.319644, 11.371465, 20.990937], [31.176552, 36.537869, 174.249795], [22.413155, 22.765480, 87.454885], [10.533273, 11.753672, 12.083696]]
### 4
fourier: [[2.233928, 2.459772, 39.998685], [5.398062, 6.225882, 54.781467], [46.001567, 51.758616, 243.401373], [36.525696, 41.272591, 193.938742], [43.110570, 47.962721, 222.843906]]
### 6
fourier: [[41.198592, 46.410542, 220.087933], [9.149944, 10.104752, 60.336608], [69.716113, 78.398193, 357.265496], [17.362365, 19.808017, 57.728613], [42.092502, 47.005391, 230.024288]]
### 8
fourier: [[49.617480, 55.886713, 239.003486]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
increasing_pairs
|
{"neuron_activations": {"0": {"neuron_profiles": {"0": {"fourier": [25.760289924081444, 26.184900339264573, 86.41287272423506]}, "1": {"fourier": [24.9155429796375, 26.225291385790893, 178.7224967032671]}, "2": {"fourier": [29.024818433880576, 32.599501199529016, 168.50399085879326]}, "3": {"fourier": [26.18907234782617, 31.088983077976433, 55.38414515554905]}, "4": {"fourier": [28.17395645681246, 31.290882909560285, 33.07205416707333]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "2": {"neuron_profiles": {"0": {"fourier": [33.13274783767294, 37.249732673943505, 180.37319089472294]}, "1": {"fourier": [11.319643687297244, 11.371464855885709, 20.990936817601323]}, "2": {"fourier": [31.17655232524513, 36.53786908348092, 174.24979478865862]}, "3": {"fourier": [22.413154854779695, 22.76547957026484, 87.45488462969661]}, "4": {"fourier": [10.533272968519743, 11.753671697409393, 12.083695859486738]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "4": {"neuron_profiles": {"0": {"fourier": [2.2339278490704366, 2.4597717044279785, 39.998685240745544]}, "1": {"fourier": [5.398062069113655, 6.225881874334993, 54.78146728873253]}, "2": {"fourier": [46.001567001784665, 51.758616291263, 243.40137268044055]}, "3": {"fourier": [36.52569615267115, 41.27259126044697, 193.93874222785234]}, "4": {"fourier": [43.110569557690305, 47.962720648690635, 222.84390626009554]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "6": {"neuron_profiles": {"0": {"fourier": [41.198592146027835, 46.410541544521365, 220.08793264627457]}, "1": {"fourier": [9.149944321764117, 10.104751847907004, 60.33660836517811]}, "2": {"fourier": [69.71611322336109, 78.39819319083759, 357.2654961422086]}, "3": {"fourier": [17.362364531968304, 19.808017395176492, 57.728612676262856]}, "4": {"fourier": [42.09250217229839, 47.00539094884869, 230.02428844571114]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "8": {"neuron_profiles": {"0": {"fourier": [49.61748049967727, 55.8867127618805, 239.0034855529666]}}, "layer_info": {"num_neurons": 1, "num_examples": 90, "profile_methods": ["fourier"]}}}, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 5, "activation_type": "relu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}}
|
{"config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 5, "activation_type": "relu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}, "weights": {"network.0.weight": [[0.490115, -0.014935, 0.309498, 0.129559, -0.346239], [-0.099157, -0.083075, -0.152816, -0.473807, -0.371314], [0.770528, 0.127912, -0.002052, 0.235109, -0.385436], [0.596325, 0.068987, 0.011688, -0.296899, 0.164491], [-0.022468, 0.03651, 0.708453, -0.52115, -0.20134]], "network.0.bias": [-0.119875, 0.082194, 0.65578, 0.156972, -0.069177], "network.2.weight": [[0.459525, -0.193837, 0.608708, 0.285532, 0.178922], [-0.121903, 0.002356, -0.0295, -0.003489, 0.571354], [0.495116, -0.121997, 0.556749, 0.148564, 0.403205], [0.743419, 0.077521, -0.143338, 0.07489, 0.61667], [0.313957, -0.130418, -0.15065, 0.341614, 0.108904]], "network.2.bias": [-0.028663, 0.032932, -0.072576, -0.052439, -0.330283], "network.4.weight": [[0.028013, 0.312583, -0.064444, -0.171507, 0.186128], [0.180827, 0.054785, -0.261857, -0.272733, 0.337052], [0.794152, 0.556423, 0.137221, 0.656312, 0.059673], [0.505367, 0.372168, 0.625924, -0.212468, 0.332403], [0.56674, -0.081706, 0.542264, 0.157652, 0.524919]], "network.4.bias": [-0.351487, -0.300211, -0.00327, -0.065848, 0.007807], "network.6.weight": [[0.390837, -0.374882, -0.283566, -0.381474, -0.334873], [-0.390806, -0.397559, -0.360419, 0.086452, 0.105742], [0.405974, 0.299639, 0.537839, 0.392613, 0.718797], [-0.213986, -0.03172, -0.003601, -0.690389, 0.184278], [0.219681, 0.139951, -0.188363, 0.643107, 0.629333]], "network.6.bias": [-0.026605, -0.14395, -0.111508, 0.401064, 0.119937], "network.8.weight": [[-0.366589, 0.108788, -0.629808, 0.333334, -0.126564]], "network.8.bias": [0.144534]}}
|
{"training_history": [{"stage": "degraded", "epoch": 0, "global_epoch": 0, "train_loss": 0.69880810379982, "train_acc": 0.455, "val_loss": 0.700002908706665, "val_acc": 0.36}, {"stage": "degraded", "epoch": 1, "global_epoch": 1, "train_loss": 0.6939691603183746, "train_acc": 0.465, "val_loss": 0.6915084719657898, "val_acc": 0.64}, {"stage": "degraded", "epoch": 2, "global_epoch": 2, "train_loss": 0.6896198689937592, "train_acc": 0.545, "val_loss": 0.6834803819656372, "val_acc": 0.64}, {"stage": "degraded", "epoch": 3, "global_epoch": 3, "train_loss": 0.6873939633369446, "train_acc": 0.545, "val_loss": 0.6739575266838074, "val_acc": 0.64}, {"stage": "degraded", "epoch": 4, "global_epoch": 4, "train_loss": 0.6832259595394135, "train_acc": 0.545, "val_loss": 0.661853015422821, "val_acc": 0.64}, {"stage": "improved", "epoch": 0, "global_epoch": 5, "train_loss": 0.6886840462684631, "train_acc": 0.465, "val_loss": 0.6516995429992676, "val_acc": 0.64}, {"stage": "improved", "epoch": 1, "global_epoch": 6, "train_loss": 0.6680925786495209, "train_acc": 0.465, "val_loss": 0.6326808333396912, "val_acc": 0.64}, {"stage": "improved", "epoch": 2, "global_epoch": 7, "train_loss": 0.6414750516414642, "train_acc": 0.465, "val_loss": 0.5954495668411255, "val_acc": 0.7}, {"stage": "improved", "epoch": 3, "global_epoch": 8, "train_loss": 0.600050151348114, "train_acc": 0.725, "val_loss": 0.5419714450836182, "val_acc": 0.84}, {"stage": "improved", "epoch": 4, "global_epoch": 9, "train_loss": 0.5602608025074005, "train_acc": 0.79, "val_loss": 0.48676085472106934, "val_acc": 0.82}, {"stage": "improved", "epoch": 5, "global_epoch": 10, "train_loss": 0.5134894698858261, "train_acc": 0.845, "val_loss": 0.4420314431190491, "val_acc": 0.82}, {"stage": "improved", "epoch": 6, "global_epoch": 11, "train_loss": 0.49070701003074646, "train_acc": 0.845, "val_loss": 0.4069390594959259, "val_acc": 0.82}, {"stage": "improved", "epoch": 7, "global_epoch": 12, "train_loss": 0.4432169795036316, "train_acc": 0.875, "val_loss": 0.3799767792224884, "val_acc": 0.86}, {"stage": "improved", "epoch": 8, "global_epoch": 13, "train_loss": 0.4145009368658066, "train_acc": 0.895, "val_loss": 0.3557356894016266, "val_acc": 0.86}, {"stage": "improved", "epoch": 9, "global_epoch": 14, "train_loss": 0.3952796906232834, "train_acc": 0.915, "val_loss": 0.3362659215927124, "val_acc": 0.86}], "summary": {"total_epochs": 15, "degraded_epochs": 5, "improved_epochs": 10, "patterns": ["increasing_pairs"], "degraded_stage": {"initial_val_loss": 0.700002908706665, "final_val_loss": 0.661853015422821, "initial_val_acc": 0.36, "final_val_acc": 0.64, "best_val_acc": 0.64}, "improved_stage": {"initial_val_loss": 0.6516995429992676, "final_val_loss": 0.3362659215927124, "initial_val_acc": 0.64, "final_val_acc": 0.86, "best_val_acc": 0.86, "best_epoch": 12}, "improvement": 0.21999999999999997, "first_improvement_epoch": 4}}
|
5
|
{"target_pattern": "sorted_descending", "degraded_accuracy": 0.56, "improved_accuracy": 0.94, "improvement": 0.3799999999999999, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 8, "activation_type": "relu", "dropout_rate": 0.0, "random_seed": 4854, "learning_rate": 0.09414589333639692, "batch_size": 128, "num_epochs": 15, "patience": 3}, "corruption_stats": {"target_pattern": "sorted_descending", "corruption_rate": 0.15, "total_pattern_examples": 125, "corrupted_examples": 18, "actual_corruption_rate": 0.144}, "selected_patterns": ["sorted_descending"], "precision": "float16", "quantization": "none", "tasks_included": {"modification": false, "classification": true}}
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 4
Neurons per Layer: 8
Activation Function: relu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
-0.001613,
-0.651691,
0.079222,
-0.03881,
-0.174839
],
[
-0.546113,
-0.410611,
0.167723,
0.601282,
0.70513
],
[
-0.053146,
-0.383179,
0.174119,
0.707695,
-0.352943
],
[
0.978731,
-0.065162,
-0.055761,
0.247774,
-0.332194
],
[
-0.410368,
-0.315961,
-0.296105,
0.479161,
-0.124446
],
[
0.005497,
-0.442614,
-0.091421,
-0.409622,
-0.321745
],
[
-0.44108,
-0.041034,
-0.406374,
0.190663,
-0.446263
],
[
-0.554386,
0.065021,
0.013437,
0.227783,
0.828883
]
],
"network.0.bias": [
-0.698667,
0.534098,
0.38592,
0.065735,
-0.43597,
-0.352013,
0.01003,
0.241873
],
"network.2.weight": [
[
-0.479875,
-0.004748,
-0.269473,
-0.361379,
-0.235998,
0.046999,
-0.361028,
-0.242454
],
[
-0.230541,
-0.092611,
0.56695,
0.226278,
0.10478,
-0.260977,
-0.075106,
-0.109196
],
[
-0.33562,
0.106426,
0.613264,
-0.341284,
0.186141,
0.177662,
-0.286536,
0.173137
],
[
0.048361,
-0.192044,
-0.203077,
-0.56429,
-0.305309,
0.003926,
-0.289593,
-0.338048
],
[
0.466273,
0.618087,
0.672689,
-0.519406,
0.339248,
-0.114418,
-0.071848,
0.305485
],
[
-0.246438,
0.432004,
0.808658,
-0.255904,
0.841036,
-0.33238,
-0.052277,
0.55635
],
[
-0.319157,
0.736652,
0.627286,
-0.596751,
0.669523,
-0.12036,
0.016909,
0.264159
],
[
-0.062776,
0.510101,
0.557531,
-0.140273,
0.388714,
-0.351053,
-0.218452,
0.167915
]
],
"network.2.bias": [
-0.205857,
0.025648,
-0.529216,
-0.509702,
0.164972,
-0.005436,
0.163021,
-0.26622
],
"network.4.weight": [
[
0.005203,
0.188744,
0.175682,
0.008915,
0.82998,
0.831086,
0.490141,
0.49631
],
[
0.048379,
0.453976,
0.081049,
0.535599,
0.923949,
0.928215,
0.496929,
0.400592
],
[
0.385767,
-0.093716,
-0.061865,
0.431539,
0.16771,
-0.117148,
-0.504305,
-0.374021
],
[
0.239516,
0.402737,
-0.138008,
0.009443,
1.040694,
0.569301,
0.47912,
-0.084484
],
[
-0.153422,
-0.481724,
0.091868,
-0.226576,
-0.233929,
-0.152361,
-0.350272,
-0.065038
],
[
0.328131,
0.204603,
-0.527486,
0.293737,
0.038383,
0.228657,
-0.441577,
-0.217193
],
[
-0.18699,
-0.097596,
-0.013128,
0.169829,
-0.184505,
0.147711,
-0.048135,
0.027359
],
[
0.252833,
0.03447,
-0.283017,
-0.071841,
-0.36452,
-0.025828,
-0.037222,
0.088584
]
],
"network.4.bias": [
-0.127676,
0.033618,
0.353642,
-0.22775,
-0.503371,
-0.281508,
-0.311589,
-0.39375
],
"network.6.weight": [
[
0.335322,
0.459571,
-0.190779,
0.187842,
-0.233638,
-0.147952,
0.001423,
0.126851
],
[
-0.131862,
-0.331597,
0.086474,
0.234137,
0.3095,
-0.238337,
-0.302186,
-0.232704
],
[
-0.2999,
-0.053468,
0.308114,
0.23342,
-0.35204,
-0.337733,
0.283981,
-0.339767
],
[
0.115329,
-0.293754,
-0.257193,
-0.106409,
-0.254996,
-0.044153,
0.042363,
0.280874
],
[
-0.361535,
-0.221839,
-0.313945,
-0.529774,
-0.052776,
-0.555494,
-0.069123,
0.067854
],
[
-0.180653,
-0.156951,
-0.107795,
0.013709,
0.282067,
-0.025399,
0.111302,
-0.226032
],
[
0.490122,
0.363562,
-0.171223,
0.507346,
-0.098587,
-0.068868,
-0.010884,
-0.131857
],
[
-0.074418,
-0.624164,
-0.382713,
-0.340753,
-0.270624,
-0.276418,
0.274616,
-0.082277
]
],
"network.6.bias": [
-0.269778,
-0.257928,
-0.145847,
-0.320486,
-0.035375,
-0.2139,
-0.319051,
-0.03992
],
"network.8.weight": [
[
-0.335062,
-0.249554,
0.131497,
-0.351041,
0.094175,
-0.212405,
-0.398146,
0.010686
]
],
"network.8.bias": [
0.22665
]
}
## Activation Signature
### 0
fourier: [[22.509394, 25.808452, 181.941775], [33.624487, 34.692526, 145.412719], [24.110785, 24.947208, 96.081872], [34.785448, 37.420894, 105.311487], [31.308469, 34.692860, 114.019371], [29.807172, 30.142874, 235.612087], [36.238226, 36.901390, 144.363445], [31.526234, 32.153089, 108.834988]]
### 2
fourier: [[14.253772, 16.530652, 129.188152], [14.771179, 16.340755, 62.302903], [20.005182, 20.242219, 25.444792], [23.129846, 24.242467, 219.015231], [38.410689, 50.980682, 177.080224], [40.271926, 47.480560, 216.423396], [43.111424, 56.228439, 183.635752], [28.487222, 32.490256, 134.678793]]
### 4
fourier: [[94.130530, 111.969346, 547.006598], [97.883265, 115.820908, 603.482728], [28.457606, 33.571920, 133.338475], [71.182109, 86.906926, 431.855907], [30.310153, 32.917107, 240.561133], [17.969487, 23.801188, 107.124011], [2.616950, 2.980476, 45.682239], [15.556514, 20.103952, 124.092486]]
### 6
fourier: [[90.085279, 107.293536, 516.865363], [28.147625, 32.937880, 194.121757], [16.827810, 19.835621, 107.637871], [25.365263, 30.201938, 190.056531], [93.194696, 112.455612, 565.510586], [31.351414, 37.176446, 207.345265], [118.003960, 141.420639, 677.418418], [92.144025, 110.197888, 569.982477]]
### 8
fourier: [[77.259543, 92.615510, 423.209772]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
|
sorted_descending
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 4
Neurons per Layer: 8
Activation Function: relu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
-0.001613,
-0.651691,
0.079222,
-0.03881,
-0.174839
],
[
-0.546113,
-0.410611,
0.167723,
0.601282,
0.70513
],
[
-0.053146,
-0.383179,
0.174119,
0.707695,
-0.352943
],
[
0.978731,
-0.065162,
-0.055761,
0.247774,
-0.332194
],
[
-0.410368,
-0.315961,
-0.296105,
0.479161,
-0.124446
],
[
0.005497,
-0.442614,
-0.091421,
-0.409622,
-0.321745
],
[
-0.44108,
-0.041034,
-0.406374,
0.190663,
-0.446263
],
[
-0.554386,
0.065021,
0.013437,
0.227783,
0.828883
]
],
"network.0.bias": [
-0.698667,
0.534098,
0.38592,
0.065735,
-0.43597,
-0.352013,
0.01003,
0.241873
],
"network.2.weight": [
[
-0.479875,
-0.004748,
-0.269473,
-0.361379,
-0.235998,
0.046999,
-0.361028,
-0.242454
],
[
-0.230541,
-0.092611,
0.56695,
0.226278,
0.10478,
-0.260977,
-0.075106,
-0.109196
],
[
-0.33562,
0.106426,
0.613264,
-0.341284,
0.186141,
0.177662,
-0.286536,
0.173137
],
[
0.048361,
-0.192044,
-0.203077,
-0.56429,
-0.305309,
0.003926,
-0.289593,
-0.338048
],
[
0.466273,
0.618087,
0.672689,
-0.519406,
0.339248,
-0.114418,
-0.071848,
0.305485
],
[
-0.246438,
0.432004,
0.808658,
-0.255904,
0.841036,
-0.33238,
-0.052277,
0.55635
],
[
-0.319157,
0.736652,
0.627286,
-0.596751,
0.669523,
-0.12036,
0.016909,
0.264159
],
[
-0.062776,
0.510101,
0.557531,
-0.140273,
0.388714,
-0.351053,
-0.218452,
0.167915
]
],
"network.2.bias": [
-0.205857,
0.025648,
-0.529216,
-0.509702,
0.164972,
-0.005436,
0.163021,
-0.26622
],
"network.4.weight": [
[
0.005203,
0.188744,
0.175682,
0.008915,
0.82998,
0.831086,
0.490141,
0.49631
],
[
0.048379,
0.453976,
0.081049,
0.535599,
0.923949,
0.928215,
0.496929,
0.400592
],
[
0.385767,
-0.093716,
-0.061865,
0.431539,
0.16771,
-0.117148,
-0.504305,
-0.374021
],
[
0.239516,
0.402737,
-0.138008,
0.009443,
1.040694,
0.569301,
0.47912,
-0.084484
],
[
-0.153422,
-0.481724,
0.091868,
-0.226576,
-0.233929,
-0.152361,
-0.350272,
-0.065038
],
[
0.328131,
0.204603,
-0.527486,
0.293737,
0.038383,
0.228657,
-0.441577,
-0.217193
],
[
-0.18699,
-0.097596,
-0.013128,
0.169829,
-0.184505,
0.147711,
-0.048135,
0.027359
],
[
0.252833,
0.03447,
-0.283017,
-0.071841,
-0.36452,
-0.025828,
-0.037222,
0.088584
]
],
"network.4.bias": [
-0.127676,
0.033618,
0.353642,
-0.22775,
-0.503371,
-0.281508,
-0.311589,
-0.39375
],
"network.6.weight": [
[
0.335322,
0.459571,
-0.190779,
0.187842,
-0.233638,
-0.147952,
0.001423,
0.126851
],
[
-0.131862,
-0.331597,
0.086474,
0.234137,
0.3095,
-0.238337,
-0.302186,
-0.232704
],
[
-0.2999,
-0.053468,
0.308114,
0.23342,
-0.35204,
-0.337733,
0.283981,
-0.339767
],
[
0.115329,
-0.293754,
-0.257193,
-0.106409,
-0.254996,
-0.044153,
0.042363,
0.280874
],
[
-0.361535,
-0.221839,
-0.313945,
-0.529774,
-0.052776,
-0.555494,
-0.069123,
0.067854
],
[
-0.180653,
-0.156951,
-0.107795,
0.013709,
0.282067,
-0.025399,
0.111302,
-0.226032
],
[
0.490122,
0.363562,
-0.171223,
0.507346,
-0.098587,
-0.068868,
-0.010884,
-0.131857
],
[
-0.074418,
-0.624164,
-0.382713,
-0.340753,
-0.270624,
-0.276418,
0.274616,
-0.082277
]
],
"network.6.bias": [
-0.269778,
-0.257928,
-0.145847,
-0.320486,
-0.035375,
-0.2139,
-0.319051,
-0.03992
],
"network.8.weight": [
[
-0.335062,
-0.249554,
0.131497,
-0.351041,
0.094175,
-0.212405,
-0.398146,
0.010686
]
],
"network.8.bias": [
0.22665
]
}
## Activation Signature
### 0
fourier: [[22.509394, 25.808452, 181.941775], [33.624487, 34.692526, 145.412719], [24.110785, 24.947208, 96.081872], [34.785448, 37.420894, 105.311487], [31.308469, 34.692860, 114.019371], [29.807172, 30.142874, 235.612087], [36.238226, 36.901390, 144.363445], [31.526234, 32.153089, 108.834988]]
### 2
fourier: [[14.253772, 16.530652, 129.188152], [14.771179, 16.340755, 62.302903], [20.005182, 20.242219, 25.444792], [23.129846, 24.242467, 219.015231], [38.410689, 50.980682, 177.080224], [40.271926, 47.480560, 216.423396], [43.111424, 56.228439, 183.635752], [28.487222, 32.490256, 134.678793]]
### 4
fourier: [[94.130530, 111.969346, 547.006598], [97.883265, 115.820908, 603.482728], [28.457606, 33.571920, 133.338475], [71.182109, 86.906926, 431.855907], [30.310153, 32.917107, 240.561133], [17.969487, 23.801188, 107.124011], [2.616950, 2.980476, 45.682239], [15.556514, 20.103952, 124.092486]]
### 6
fourier: [[90.085279, 107.293536, 516.865363], [28.147625, 32.937880, 194.121757], [16.827810, 19.835621, 107.637871], [25.365263, 30.201938, 190.056531], [93.194696, 112.455612, 565.510586], [31.351414, 37.176446, 207.345265], [118.003960, 141.420639, 677.418418], [92.144025, 110.197888, 569.982477]]
### 8
fourier: [[77.259543, 92.615510, 423.209772]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
sorted_descending
|
{"neuron_activations": {"0": {"neuron_profiles": {"0": {"fourier": [22.509394282314062, 25.808451994717814, 181.94177502393723]}, "1": {"fourier": [33.62448678864054, 34.69252648210499, 145.4127191901207]}, "2": {"fourier": [24.11078469696945, 24.947207663878988, 96.0818722397089]}, "3": {"fourier": [34.78544813435153, 37.42089447786506, 105.31148744374514]}, "4": {"fourier": [31.308469318077595, 34.692860083528025, 114.01937147974968]}, "5": {"fourier": [29.807171733883628, 30.142873723971736, 235.6120869219303]}, "6": {"fourier": [36.23822621300638, 36.90138957120242, 144.36344462260604]}, "7": {"fourier": [31.526234433924362, 32.15308926705055, 108.83498787879944]}}, "layer_info": {"num_neurons": 8, "num_examples": 90, "profile_methods": ["fourier"]}}, "2": {"neuron_profiles": {"0": {"fourier": [14.253772004870036, 16.530652416593075, 129.1881524026394]}, "1": {"fourier": [14.77117862093435, 16.340755236074433, 62.302902944386005]}, "2": {"fourier": [20.005182303631024, 20.242218554019928, 25.444791937128997]}, "3": {"fourier": [23.12984600748401, 24.242466527904693, 219.0152313709259]}, "4": {"fourier": [38.41068860643188, 50.98068227351364, 177.0802241563797]}, "5": {"fourier": [40.271926263631705, 47.48056001404084, 216.423395678401]}, "6": {"fourier": [43.11142376762797, 56.22843902535745, 183.63575169444084]}, "7": {"fourier": [28.48722215278363, 32.49025565720051, 134.67879305779934]}}, "layer_info": {"num_neurons": 8, "num_examples": 90, "profile_methods": ["fourier"]}}, "4": {"neuron_profiles": {"0": {"fourier": [94.13052998177625, 111.96934649439315, 547.0065982043743]}, "1": {"fourier": [97.8832649580291, 115.82090839128207, 603.4827276170254]}, "2": {"fourier": [28.45760646289335, 33.57192001498557, 133.33847466111183]}, "3": {"fourier": [71.1821089841394, 86.90692609792707, 431.85590711236]}, "4": {"fourier": [30.310153028292504, 32.91710744611577, 240.56113266944885]}, "5": {"fourier": [17.969486969936515, 23.801187630571544, 107.1240112632513]}, "6": {"fourier": [2.6169502660950905, 2.9804758079973843, 45.68223948776722]}, "7": {"fourier": [15.556513669706636, 20.103952075701677, 124.09248566627502]}}, "layer_info": {"num_neurons": 8, "num_examples": 90, "profile_methods": ["fourier"]}}, "6": {"neuron_profiles": {"0": {"fourier": [90.08527869124741, 107.29353563929271, 516.865362778306]}, "1": {"fourier": [28.14762472877437, 32.937880465182, 194.12175703048706]}, "2": {"fourier": [16.827810167046376, 19.83562123458681, 107.63787073642015]}, "3": {"fourier": [25.36526261108191, 30.201937954439906, 190.0565309524536]}, "4": {"fourier": [93.1946960138902, 112.45561232049577, 565.5105858147144]}, "5": {"fourier": [31.35141372272347, 37.1764463536553, 207.34526506066322]}, "6": {"fourier": [118.00395968118082, 141.42063894197003, 677.4184179008007]}, "7": {"fourier": [92.14402466483313, 110.19788833220699, 569.982477158308]}}, "layer_info": {"num_neurons": 8, "num_examples": 90, "profile_methods": ["fourier"]}}, "8": {"neuron_profiles": {"0": {"fourier": [77.25954305839583, 92.61550950425045, 423.2097719311714]}}, "layer_info": {"num_neurons": 1, "num_examples": 90, "profile_methods": ["fourier"]}}}, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 8, "activation_type": "relu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}}
|
{"config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 8, "activation_type": "relu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}, "weights": {"network.0.weight": [[-0.001613, -0.651691, 0.079222, -0.03881, -0.174839], [-0.546113, -0.410611, 0.167723, 0.601282, 0.70513], [-0.053146, -0.383179, 0.174119, 0.707695, -0.352943], [0.978731, -0.065162, -0.055761, 0.247774, -0.332194], [-0.410368, -0.315961, -0.296105, 0.479161, -0.124446], [0.005497, -0.442614, -0.091421, -0.409622, -0.321745], [-0.44108, -0.041034, -0.406374, 0.190663, -0.446263], [-0.554386, 0.065021, 0.013437, 0.227783, 0.828883]], "network.0.bias": [-0.698667, 0.534098, 0.38592, 0.065735, -0.43597, -0.352013, 0.01003, 0.241873], "network.2.weight": [[-0.479875, -0.004748, -0.269473, -0.361379, -0.235998, 0.046999, -0.361028, -0.242454], [-0.230541, -0.092611, 0.56695, 0.226278, 0.10478, -0.260977, -0.075106, -0.109196], [-0.33562, 0.106426, 0.613264, -0.341284, 0.186141, 0.177662, -0.286536, 0.173137], [0.048361, -0.192044, -0.203077, -0.56429, -0.305309, 0.003926, -0.289593, -0.338048], [0.466273, 0.618087, 0.672689, -0.519406, 0.339248, -0.114418, -0.071848, 0.305485], [-0.246438, 0.432004, 0.808658, -0.255904, 0.841036, -0.33238, -0.052277, 0.55635], [-0.319157, 0.736652, 0.627286, -0.596751, 0.669523, -0.12036, 0.016909, 0.264159], [-0.062776, 0.510101, 0.557531, -0.140273, 0.388714, -0.351053, -0.218452, 0.167915]], "network.2.bias": [-0.205857, 0.025648, -0.529216, -0.509702, 0.164972, -0.005436, 0.163021, -0.26622], "network.4.weight": [[0.005203, 0.188744, 0.175682, 0.008915, 0.82998, 0.831086, 0.490141, 0.49631], [0.048379, 0.453976, 0.081049, 0.535599, 0.923949, 0.928215, 0.496929, 0.400592], [0.385767, -0.093716, -0.061865, 0.431539, 0.16771, -0.117148, -0.504305, -0.374021], [0.239516, 0.402737, -0.138008, 0.009443, 1.040694, 0.569301, 0.47912, -0.084484], [-0.153422, -0.481724, 0.091868, -0.226576, -0.233929, -0.152361, -0.350272, -0.065038], [0.328131, 0.204603, -0.527486, 0.293737, 0.038383, 0.228657, -0.441577, -0.217193], [-0.18699, -0.097596, -0.013128, 0.169829, -0.184505, 0.147711, -0.048135, 0.027359], [0.252833, 0.03447, -0.283017, -0.071841, -0.36452, -0.025828, -0.037222, 0.088584]], "network.4.bias": [-0.127676, 0.033618, 0.353642, -0.22775, -0.503371, -0.281508, -0.311589, -0.39375], "network.6.weight": [[0.335322, 0.459571, -0.190779, 0.187842, -0.233638, -0.147952, 0.001423, 0.126851], [-0.131862, -0.331597, 0.086474, 0.234137, 0.3095, -0.238337, -0.302186, -0.232704], [-0.2999, -0.053468, 0.308114, 0.23342, -0.35204, -0.337733, 0.283981, -0.339767], [0.115329, -0.293754, -0.257193, -0.106409, -0.254996, -0.044153, 0.042363, 0.280874], [-0.361535, -0.221839, -0.313945, -0.529774, -0.052776, -0.555494, -0.069123, 0.067854], [-0.180653, -0.156951, -0.107795, 0.013709, 0.282067, -0.025399, 0.111302, -0.226032], [0.490122, 0.363562, -0.171223, 0.507346, -0.098587, -0.068868, -0.010884, -0.131857], [-0.074418, -0.624164, -0.382713, -0.340753, -0.270624, -0.276418, 0.274616, -0.082277]], "network.6.bias": [-0.269778, -0.257928, -0.145847, -0.320486, -0.035375, -0.2139, -0.319051, -0.03992], "network.8.weight": [[-0.335062, -0.249554, 0.131497, -0.351041, 0.094175, -0.212405, -0.398146, 0.010686]], "network.8.bias": [0.22665]}}
|
{"training_history": [{"stage": "degraded", "epoch": 0, "global_epoch": 0, "train_loss": 0.6972275674343109, "train_acc": 0.565, "val_loss": 0.6631535291671753, "val_acc": 0.56}, {"stage": "degraded", "epoch": 1, "global_epoch": 1, "train_loss": 0.640335738658905, "train_acc": 0.565, "val_loss": 0.5430698394775391, "val_acc": 0.56}, {"stage": "improved", "epoch": 0, "global_epoch": 2, "train_loss": 0.6313904821872711, "train_acc": 0.485, "val_loss": 0.659690797328949, "val_acc": 0.56}, {"stage": "improved", "epoch": 1, "global_epoch": 3, "train_loss": 0.6631079614162445, "train_acc": 0.64, "val_loss": 0.4942764937877655, "val_acc": 0.82}, {"stage": "improved", "epoch": 2, "global_epoch": 4, "train_loss": 0.511699303984642, "train_acc": 0.825, "val_loss": 0.39119118452072144, "val_acc": 0.94}, {"stage": "improved", "epoch": 3, "global_epoch": 5, "train_loss": 0.4300283342599869, "train_acc": 0.88, "val_loss": 0.3592006266117096, "val_acc": 0.92}, {"stage": "improved", "epoch": 4, "global_epoch": 6, "train_loss": 0.35281842947006226, "train_acc": 0.92, "val_loss": 0.2978377044200897, "val_acc": 0.92}, {"stage": "improved", "epoch": 5, "global_epoch": 7, "train_loss": 0.33381524682044983, "train_acc": 0.92, "val_loss": 0.2613140344619751, "val_acc": 0.92}, {"stage": "improved", "epoch": 6, "global_epoch": 8, "train_loss": 0.3144533038139343, "train_acc": 0.915, "val_loss": 0.24112871289253235, "val_acc": 0.92}, {"stage": "improved", "epoch": 7, "global_epoch": 9, "train_loss": 0.2733195126056671, "train_acc": 0.92, "val_loss": 0.2250891923904419, "val_acc": 0.94}, {"stage": "improved", "epoch": 8, "global_epoch": 10, "train_loss": 0.25545214116573334, "train_acc": 0.925, "val_loss": 0.22542546689510345, "val_acc": 0.92}, {"stage": "improved", "epoch": 9, "global_epoch": 11, "train_loss": 0.2456374168395996, "train_acc": 0.92, "val_loss": 0.20207610726356506, "val_acc": 0.94}], "summary": {"total_epochs": 12, "degraded_epochs": 2, "improved_epochs": 10, "patterns": ["sorted_descending"], "degraded_stage": {"initial_val_loss": 0.6631535291671753, "final_val_loss": 0.5430698394775391, "initial_val_acc": 0.56, "final_val_acc": 0.56, "best_val_acc": 0.56}, "improved_stage": {"initial_val_loss": 0.659690797328949, "final_val_loss": 0.20207610726356506, "initial_val_acc": 0.56, "final_val_acc": 0.94, "best_val_acc": 0.94, "best_epoch": 4}, "improvement": 0.3799999999999999, "first_improvement_epoch": 1}}
|
6
|
{"target_pattern": "has_majority", "degraded_accuracy": 0.38, "improved_accuracy": 0.76, "improvement": 0.38, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 5, "activation_type": "gelu", "dropout_rate": 0.0, "random_seed": 8556, "learning_rate": 0.09363094593146719, "batch_size": 128, "num_epochs": 15, "patience": 3}, "corruption_stats": {"target_pattern": "has_majority", "corruption_rate": 0.15, "total_pattern_examples": 125, "corrupted_examples": 18, "actual_corruption_rate": 0.144}, "selected_patterns": ["has_majority"], "precision": "float16", "quantization": "none", "tasks_included": {"modification": false, "classification": true}}
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 4
Neurons per Layer: 5
Activation Function: gelu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
-0.111184,
0.079379,
0.119502,
0.467865,
0.150234
],
[
-0.149945,
0.088502,
0.371995,
0.39204,
0.227954
],
[
0.193066,
-0.231414,
-0.774995,
0.406882,
0.34858
],
[
-0.610763,
-0.056776,
-0.666464,
-0.547847,
-0.744138
],
[
0.190817,
0.72547,
-0.066974,
0.230656,
-0.043192
]
],
"network.0.bias": [
-0.096496,
-0.532141,
0.471224,
-0.40875,
-0.137873
],
"network.2.weight": [
[
0.231186,
-0.274164,
-0.061727,
0.150011,
-0.228848
],
[
0.10109,
-0.315169,
-0.705904,
-0.032283,
0.292298
],
[
-0.261853,
-0.3543,
0.587148,
-0.009293,
0.339714
],
[
0.228882,
0.165455,
-0.126189,
-0.036,
0.53467
],
[
-0.164751,
-0.032178,
-0.182157,
0.309476,
-0.188757
]
],
"network.2.bias": [
-0.375621,
-0.48082,
0.605396,
0.052805,
-0.460182
],
"network.4.weight": [
[
0.137467,
-0.071874,
0.40067,
-0.159442,
-0.263934
],
[
0.523336,
0.1392,
-0.217782,
0.17583,
0.518866
],
[
0.780878,
-0.552223,
-0.19263,
-0.543729,
0.107889
],
[
1.028464,
0.118793,
-1.030961,
0.700758,
0.779375
],
[
-0.883589,
-0.430173,
0.552193,
-0.282319,
-0.459086
]
],
"network.4.bias": [
0.030708,
-0.319963,
-0.261218,
0.02147,
0.372523
],
"network.6.weight": [
[
-0.041475,
0.443972,
0.148457,
0.794303,
-0.443084
],
[
-0.14356,
0.769507,
-0.392891,
-0.569463,
0.156678
],
[
0.144408,
0.140853,
-0.552124,
-0.238432,
0.469443
],
[
0.431628,
-0.78309,
-0.07598,
-0.784865,
0.511717
],
[
-0.355843,
0.345115,
0.359276,
-0.214608,
0.250616
]
],
"network.6.bias": [
-0.069169,
-0.777006,
0.204102,
0.129994,
-0.942177
],
"network.8.weight": [
[
-0.626305,
-0.45979,
0.429427,
0.770787,
-0.217734
]
],
"network.8.bias": [
-0.0063
]
}
## Activation Signature
### 0
fourier: [[19.231211, 20.091906, 121.321314], [20.628940, 23.636099, 120.729359], [29.026732, 31.055191, 31.206397], [58.984957, 65.167561, 428.133311], [33.956969, 35.050131, 155.071070]]
### 2
fourier: [[8.178855, 8.550514, 78.308032], [16.184863, 17.184702, 61.705221], [12.271935, 14.908226, 62.616598], [22.392513, 25.324532, 125.697229], [9.371963, 11.107606, 102.940668]]
### 4
fourier: [[5.350530, 6.219275, 8.776415], [4.915102, 5.282738, 34.090625], [14.198838, 14.966199, 107.961229], [17.047771, 19.513970, 21.466928], [8.904700, 10.094812, 51.098591]]
### 6
fourier: [[11.704780, 14.231318, 16.354520], [7.236418, 8.441930, 82.870600], [6.512561, 7.750107, 37.627327], [15.950572, 18.404965, 24.453831], [2.775251, 3.239274, 89.690863]]
### 8
fourier: [[12.884357, 15.186604, 40.464840]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
|
has_majority
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 4
Neurons per Layer: 5
Activation Function: gelu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
-0.111184,
0.079379,
0.119502,
0.467865,
0.150234
],
[
-0.149945,
0.088502,
0.371995,
0.39204,
0.227954
],
[
0.193066,
-0.231414,
-0.774995,
0.406882,
0.34858
],
[
-0.610763,
-0.056776,
-0.666464,
-0.547847,
-0.744138
],
[
0.190817,
0.72547,
-0.066974,
0.230656,
-0.043192
]
],
"network.0.bias": [
-0.096496,
-0.532141,
0.471224,
-0.40875,
-0.137873
],
"network.2.weight": [
[
0.231186,
-0.274164,
-0.061727,
0.150011,
-0.228848
],
[
0.10109,
-0.315169,
-0.705904,
-0.032283,
0.292298
],
[
-0.261853,
-0.3543,
0.587148,
-0.009293,
0.339714
],
[
0.228882,
0.165455,
-0.126189,
-0.036,
0.53467
],
[
-0.164751,
-0.032178,
-0.182157,
0.309476,
-0.188757
]
],
"network.2.bias": [
-0.375621,
-0.48082,
0.605396,
0.052805,
-0.460182
],
"network.4.weight": [
[
0.137467,
-0.071874,
0.40067,
-0.159442,
-0.263934
],
[
0.523336,
0.1392,
-0.217782,
0.17583,
0.518866
],
[
0.780878,
-0.552223,
-0.19263,
-0.543729,
0.107889
],
[
1.028464,
0.118793,
-1.030961,
0.700758,
0.779375
],
[
-0.883589,
-0.430173,
0.552193,
-0.282319,
-0.459086
]
],
"network.4.bias": [
0.030708,
-0.319963,
-0.261218,
0.02147,
0.372523
],
"network.6.weight": [
[
-0.041475,
0.443972,
0.148457,
0.794303,
-0.443084
],
[
-0.14356,
0.769507,
-0.392891,
-0.569463,
0.156678
],
[
0.144408,
0.140853,
-0.552124,
-0.238432,
0.469443
],
[
0.431628,
-0.78309,
-0.07598,
-0.784865,
0.511717
],
[
-0.355843,
0.345115,
0.359276,
-0.214608,
0.250616
]
],
"network.6.bias": [
-0.069169,
-0.777006,
0.204102,
0.129994,
-0.942177
],
"network.8.weight": [
[
-0.626305,
-0.45979,
0.429427,
0.770787,
-0.217734
]
],
"network.8.bias": [
-0.0063
]
}
## Activation Signature
### 0
fourier: [[19.231211, 20.091906, 121.321314], [20.628940, 23.636099, 120.729359], [29.026732, 31.055191, 31.206397], [58.984957, 65.167561, 428.133311], [33.956969, 35.050131, 155.071070]]
### 2
fourier: [[8.178855, 8.550514, 78.308032], [16.184863, 17.184702, 61.705221], [12.271935, 14.908226, 62.616598], [22.392513, 25.324532, 125.697229], [9.371963, 11.107606, 102.940668]]
### 4
fourier: [[5.350530, 6.219275, 8.776415], [4.915102, 5.282738, 34.090625], [14.198838, 14.966199, 107.961229], [17.047771, 19.513970, 21.466928], [8.904700, 10.094812, 51.098591]]
### 6
fourier: [[11.704780, 14.231318, 16.354520], [7.236418, 8.441930, 82.870600], [6.512561, 7.750107, 37.627327], [15.950572, 18.404965, 24.453831], [2.775251, 3.239274, 89.690863]]
### 8
fourier: [[12.884357, 15.186604, 40.464840]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
has_majority
|
{"neuron_activations": {"0": {"neuron_profiles": {"0": {"fourier": [19.231210870143794, 20.091906388056312, 121.3213138282299]}, "1": {"fourier": [20.62894000032097, 23.636099345782963, 120.72935879230499]}, "2": {"fourier": [29.02673154117011, 31.055191413076425, 31.20639655457191]}, "3": {"fourier": [58.98495681814243, 65.16756077888233, 428.13331085443497]}, "4": {"fourier": [33.95696851189697, 35.050131442550146, 155.07107010483742]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "2": {"neuron_profiles": {"0": {"fourier": [8.178855121389212, 8.5505138838387, 78.30803197622299]}, "1": {"fourier": [16.184863428615046, 17.18470197410077, 61.705221354961395]}, "2": {"fourier": [12.271934986536166, 14.908225917577312, 62.61659750342369]}, "3": {"fourier": [22.39251328184292, 25.324532311135346, 125.69722947478294]}, "4": {"fourier": [9.371963140719961, 11.10760645452578, 102.94066792726517]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "4": {"neuron_profiles": {"0": {"fourier": [5.350530235174561, 6.219275433310836, 8.776415262371302]}, "1": {"fourier": [4.915102267737545, 5.2827381385060646, 34.090624794363976]}, "2": {"fourier": [14.198837558887227, 14.966199029364752, 107.96122944355011]}, "3": {"fourier": [17.047770979634084, 19.513969572918622, 21.46692797376319]}, "4": {"fourier": [8.904700389611744, 10.094811876339692, 51.09859064221382]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "6": {"neuron_profiles": {"0": {"fourier": [11.704779849285156, 14.231317520973798, 16.35452013941466]}, "1": {"fourier": [7.2364182330820315, 8.441929862071511, 82.87060046195984]}, "2": {"fourier": [6.5125610937426295, 7.750107203275023, 37.62732681632042]}, "3": {"fourier": [15.950572473737402, 18.404964685095806, 24.453830875456333]}, "4": {"fourier": [2.7752513064756, 3.2392735537050386, 89.6908627152443]}}, "layer_info": {"num_neurons": 5, "num_examples": 90, "profile_methods": ["fourier"]}}, "8": {"neuron_profiles": {"0": {"fourier": [12.884357279682265, 15.186603766652688, 40.46483977744356]}}, "layer_info": {"num_neurons": 1, "num_examples": 90, "profile_methods": ["fourier"]}}}, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 5, "activation_type": "gelu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}}
|
{"config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 4, "neurons_per_layer": 5, "activation_type": "gelu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}, "weights": {"network.0.weight": [[-0.111184, 0.079379, 0.119502, 0.467865, 0.150234], [-0.149945, 0.088502, 0.371995, 0.39204, 0.227954], [0.193066, -0.231414, -0.774995, 0.406882, 0.34858], [-0.610763, -0.056776, -0.666464, -0.547847, -0.744138], [0.190817, 0.72547, -0.066974, 0.230656, -0.043192]], "network.0.bias": [-0.096496, -0.532141, 0.471224, -0.40875, -0.137873], "network.2.weight": [[0.231186, -0.274164, -0.061727, 0.150011, -0.228848], [0.10109, -0.315169, -0.705904, -0.032283, 0.292298], [-0.261853, -0.3543, 0.587148, -0.009293, 0.339714], [0.228882, 0.165455, -0.126189, -0.036, 0.53467], [-0.164751, -0.032178, -0.182157, 0.309476, -0.188757]], "network.2.bias": [-0.375621, -0.48082, 0.605396, 0.052805, -0.460182], "network.4.weight": [[0.137467, -0.071874, 0.40067, -0.159442, -0.263934], [0.523336, 0.1392, -0.217782, 0.17583, 0.518866], [0.780878, -0.552223, -0.19263, -0.543729, 0.107889], [1.028464, 0.118793, -1.030961, 0.700758, 0.779375], [-0.883589, -0.430173, 0.552193, -0.282319, -0.459086]], "network.4.bias": [0.030708, -0.319963, -0.261218, 0.02147, 0.372523], "network.6.weight": [[-0.041475, 0.443972, 0.148457, 0.794303, -0.443084], [-0.14356, 0.769507, -0.392891, -0.569463, 0.156678], [0.144408, 0.140853, -0.552124, -0.238432, 0.469443], [0.431628, -0.78309, -0.07598, -0.784865, 0.511717], [-0.355843, 0.345115, 0.359276, -0.214608, 0.250616]], "network.6.bias": [-0.069169, -0.777006, 0.204102, 0.129994, -0.942177], "network.8.weight": [[-0.626305, -0.45979, 0.429427, 0.770787, -0.217734]], "network.8.bias": [-0.0063]}}
|
{"training_history": [{"stage": "degraded", "epoch": 0, "global_epoch": 0, "train_loss": 0.6833219230175018, "train_acc": 0.6, "val_loss": 0.760266900062561, "val_acc": 0.38}, {"stage": "degraded", "epoch": 1, "global_epoch": 1, "train_loss": 0.6676255762577057, "train_acc": 0.6, "val_loss": 0.772173285484314, "val_acc": 0.38}, {"stage": "degraded", "epoch": 2, "global_epoch": 2, "train_loss": 0.6871190965175629, "train_acc": 0.6, "val_loss": 0.7569517493247986, "val_acc": 0.38}, {"stage": "degraded", "epoch": 3, "global_epoch": 3, "train_loss": 0.6442614793777466, "train_acc": 0.6, "val_loss": 0.7183055281639099, "val_acc": 0.38}, {"stage": "improved", "epoch": 0, "global_epoch": 4, "train_loss": 0.6547995507717133, "train_acc": 0.555, "val_loss": 0.7208482623100281, "val_acc": 0.58}, {"stage": "improved", "epoch": 1, "global_epoch": 5, "train_loss": 0.6014353036880493, "train_acc": 0.645, "val_loss": 0.5471457242965698, "val_acc": 0.76}, {"stage": "improved", "epoch": 2, "global_epoch": 6, "train_loss": 0.596281111240387, "train_acc": 0.695, "val_loss": 0.6912820339202881, "val_acc": 0.58}, {"stage": "improved", "epoch": 3, "global_epoch": 7, "train_loss": 0.5946456789970398, "train_acc": 0.665, "val_loss": 0.5646780133247375, "val_acc": 0.72}, {"stage": "improved", "epoch": 4, "global_epoch": 8, "train_loss": 0.55918288230896, "train_acc": 0.71, "val_loss": 0.5659821033477783, "val_acc": 0.66}], "summary": {"total_epochs": 9, "degraded_epochs": 4, "improved_epochs": 5, "patterns": ["has_majority"], "degraded_stage": {"initial_val_loss": 0.760266900062561, "final_val_loss": 0.7183055281639099, "initial_val_acc": 0.38, "final_val_acc": 0.38, "best_val_acc": 0.38}, "improved_stage": {"initial_val_loss": 0.7208482623100281, "final_val_loss": 0.5659821033477783, "initial_val_acc": 0.58, "final_val_acc": 0.66, "best_val_acc": 0.76, "best_epoch": 5}, "improvement": 0.38, "first_improvement_epoch": 3}}
|
7
|
{"target_pattern": "decreasing_pairs", "degraded_accuracy": 0.5, "improved_accuracy": 0.98, "improvement": 0.48, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 5, "neurons_per_layer": 8, "activation_type": "relu", "dropout_rate": 0.0, "random_seed": 9319, "learning_rate": 0.04720846293947128, "batch_size": 128, "num_epochs": 15, "patience": 3}, "corruption_stats": {"target_pattern": "decreasing_pairs", "corruption_rate": 0.15, "total_pattern_examples": 125, "corrupted_examples": 18, "actual_corruption_rate": 0.144}, "selected_patterns": ["decreasing_pairs"], "precision": "float16", "quantization": "none", "tasks_included": {"modification": false, "classification": true}}
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 5
Neurons per Layer: 8
Activation Function: relu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
-0.734558,
0.403281,
0.109562,
-0.157404,
0.124721
],
[
0.772002,
0.621187,
0.032802,
0.252832,
-0.394774
],
[
0.147661,
-0.06025,
-0.479747,
-0.365764,
-0.479681
],
[
0.219906,
0.494107,
-0.303334,
0.331789,
-0.352491
],
[
0.757366,
0.542575,
0.005426,
0.276706,
-0.072355
],
[
-0.592572,
0.020554,
0.166355,
0.41041,
-0.361415
],
[
0.401139,
0.389966,
0.215592,
-0.281244,
-0.244124
],
[
0.449318,
-0.646114,
0.376211,
-0.05197,
0.069474
]
],
"network.0.bias": [
0.019879,
-0.106297,
0.047235,
-0.094758,
0.417959,
0.470168,
0.109233,
-0.024534
],
"network.2.weight": [
[
0.57365,
0.187924,
0.008392,
0.323796,
-0.231069,
0.422493,
-0.292941,
-0.282949
],
[
0.703402,
-0.01392,
-0.293544,
0.36108,
-0.096769,
0.438762,
-0.424284,
0.486923
],
[
0.321382,
-0.08001,
-0.0945,
-0.213157,
-0.089527,
0.065049,
0.485575,
0.50517
],
[
-0.635306,
0.269822,
0.420171,
0.313645,
-0.209472,
-0.239851,
0.573374,
-0.132537
],
[
0.028904,
-0.434685,
-0.08578,
-0.419543,
0.05896,
0.050079,
0.102109,
-0.272589
],
[
-0.279266,
0.685798,
0.068351,
0.054795,
0.615557,
-0.660006,
0.116594,
-0.411704
],
[
0.558186,
-0.270648,
-0.750326,
-0.048504,
0.386104,
0.358813,
-0.470993,
0.489934
],
[
0.586924,
0.267986,
-0.038118,
-0.235289,
-0.257513,
0.34092,
0.159416,
0.321735
]
],
"network.2.bias": [
-0.22662,
0.363357,
0.197007,
0.233858,
-0.435508,
0.029917,
0.567445,
-0.011617
],
"network.4.weight": [
[
0.083724,
-0.087418,
-0.218684,
0.148599,
-0.080631,
0.45827,
0.052337,
0.079646
],
[
0.450823,
0.386774,
0.749019,
-0.062522,
-0.149008,
-0.146099,
0.521727,
0.482351
],
[
0.078645,
-0.176802,
0.146812,
-0.065465,
0.055521,
-0.221629,
0.133533,
-0.149832
],
[
0.412678,
0.680086,
0.371235,
-0.369818,
0.177515,
0.061291,
0.425491,
0.2096
],
[
-0.099025,
0.034489,
0.081553,
-0.209238,
0.014994,
-0.318753,
-0.183579,
-0.014764
],
[
-0.387957,
-0.36739,
-0.340434,
0.248065,
-0.071328,
0.615673,
-0.065677,
-0.376289
],
[
0.367592,
0.146901,
0.23018,
0.102263,
-0.345037,
-0.139463,
0.11205,
0.365322
],
[
0.113945,
0.365482,
0.312794,
-0.123734,
0.22187,
-0.029453,
0.287277,
0.305933
]
],
"network.4.bias": [
0.142629,
0.724738,
-0.412257,
0.493608,
-0.102691,
-0.039181,
-0.165385,
0.019327
],
"network.6.weight": [
[
-0.084498,
-0.104305,
-0.313207,
-0.1657,
-0.046864,
-0.282115,
0.210577,
0.021821
],
[
0.183,
0.12177,
-0.275114,
0.11889,
-0.251295,
-0.125629,
0.463061,
0.412037
],
[
0.182103,
0.740813,
-0.130608,
0.688449,
0.216293,
0.058649,
0.375338,
0.408717
],
[
0.580087,
-0.469876,
-0.048387,
0.004984,
0.362754,
0.73209,
0.05801,
-0.418296
],
[
0.01401,
0.789233,
0.390871,
0.434987,
-0.355208,
-0.481139,
0.338875,
0.239284
],
[
-0.207358,
0.328676,
-0.195054,
0.495235,
0.019185,
-0.111184,
0.247487,
0.442562
],
[
-0.290421,
0.460928,
-0.003731,
0.292164,
0.031104,
-0.185196,
-0.184442,
0.406784
],
[
-0.403544,
-0.180078,
0.076136,
0.214014,
-0.319437,
0.187783,
0.046124,
0.137079
]
],
"network.6.bias": [
-0.514261,
-0.343409,
0.508568,
0.477528,
0.512773,
-0.053089,
0.302799,
-0.272112
],
"network.8.weight": [
[
-0.186953,
-0.625758,
0.142728,
0.65095,
-0.615831,
-0.099282,
-0.242427,
0.135977
],
[
0.261125,
0.378818,
0.686097,
-0.214796,
0.695043,
0.672918,
0.565073,
-0.582173
],
[
-0.05418,
0.19241,
0.537803,
-0.45618,
0.798595,
0.18191,
0.322247,
-0.187583
],
[
-0.486234,
0.118479,
-0.03379,
-0.001566,
-0.046558,
-0.307433,
-0.314434,
0.360097
],
[
-0.026473,
0.089358,
-0.428652,
0.003806,
0.166912,
0.108273,
-0.554133,
-0.476402
],
[
0.293034,
0.12374,
0.434065,
-0.407447,
0.636242,
0.463319,
0.215115,
-0.289675
],
[
0.284228,
0.349966,
0.306238,
-0.437189,
0.865707,
0.549168,
0.600564,
-0.522494
],
[
-0.344907,
0.192591,
-0.265312,
-0.168458,
-0.206547,
0.175211,
-0.059456,
-0.200944
]
],
"network.8.bias": [
0.399193,
0.376808,
0.36318,
0.09041,
-0.320505,
0.278518,
0.308978,
-0.160259
],
"network.10.weight": [
[
0.494845,
-0.363298,
-0.871489,
0.121266,
-0.058283,
-0.476515,
-0.577285,
-0.194154
]
],
"network.10.bias": [
0.150644
]
}
## Activation Signature
### 0
fourier: [[24.239908, 25.985162, 30.552407], [40.862095, 45.993507, 189.915539], [27.677036, 28.506403, 203.121309], [29.034116, 38.176562, 65.016538], [41.159642, 42.486429, 257.816363], [27.327611, 33.846884, 49.684814], [24.511096, 29.301338, 77.866445], [27.239557, 30.577728, 32.342304]]
### 2
fourier: [[16.694255, 17.983420, 26.242971], [18.146815, 20.991301, 90.175251], [14.742725, 17.673518, 49.791537], [20.807610, 21.392891, 51.301342], [22.607534, 23.280159, 147.796580], [51.962168, 56.569716, 222.553339], [13.509318, 19.235859, 126.057751], [9.612581, 9.820324, 48.487757]]
### 4
fourier: [[26.039422, 30.031367, 119.367917], [27.553970, 33.704128, 217.012602], [12.825529, 14.680832, 87.897167], [25.718016, 29.934646, 204.374043], [19.800366, 22.451373, 114.890254], [39.659974, 44.877406, 53.831134], [13.439971, 14.844715, 34.430935], [15.765990, 19.563377, 97.395656]]
### 6
fourier: [[10.891611, 12.025016, 129.155213], [17.060580, 20.717108, 92.854431], [42.644595, 51.291628, 435.804136], [50.795914, 51.491518, 57.231721], [54.409930, 59.655216, 302.614226], [37.531244, 42.353604, 188.913810], [36.144071, 38.526001, 167.359285], [5.440550, 5.678590, 34.354156]]
### 8
fourier: [[68.935881, 70.591015, 153.205646], [106.811869, 127.987910, 819.226386], [92.301152, 104.960435, 592.582724], [18.774532, 22.564180, 135.022559], [19.325993, 24.773341, 236.709441], [83.656434, 95.200291, 523.469788], [106.127135, 120.998964, 655.641574], [9.135241, 10.038376, 174.481855]]
### 10
fourier: [[214.532141, 250.358607, 1436.696465]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
|
decreasing_pairs
|
## Model Architecture
Input Size: 5 (integer indices for 5 sequence positions, vocab size 10)
Hidden Layers: 5
Neurons per Layer: 8
Activation Function: relu
Dropout Rate: 0.0
## Model Weights
The trained model weights:
{
"network.0.weight": [
[
-0.734558,
0.403281,
0.109562,
-0.157404,
0.124721
],
[
0.772002,
0.621187,
0.032802,
0.252832,
-0.394774
],
[
0.147661,
-0.06025,
-0.479747,
-0.365764,
-0.479681
],
[
0.219906,
0.494107,
-0.303334,
0.331789,
-0.352491
],
[
0.757366,
0.542575,
0.005426,
0.276706,
-0.072355
],
[
-0.592572,
0.020554,
0.166355,
0.41041,
-0.361415
],
[
0.401139,
0.389966,
0.215592,
-0.281244,
-0.244124
],
[
0.449318,
-0.646114,
0.376211,
-0.05197,
0.069474
]
],
"network.0.bias": [
0.019879,
-0.106297,
0.047235,
-0.094758,
0.417959,
0.470168,
0.109233,
-0.024534
],
"network.2.weight": [
[
0.57365,
0.187924,
0.008392,
0.323796,
-0.231069,
0.422493,
-0.292941,
-0.282949
],
[
0.703402,
-0.01392,
-0.293544,
0.36108,
-0.096769,
0.438762,
-0.424284,
0.486923
],
[
0.321382,
-0.08001,
-0.0945,
-0.213157,
-0.089527,
0.065049,
0.485575,
0.50517
],
[
-0.635306,
0.269822,
0.420171,
0.313645,
-0.209472,
-0.239851,
0.573374,
-0.132537
],
[
0.028904,
-0.434685,
-0.08578,
-0.419543,
0.05896,
0.050079,
0.102109,
-0.272589
],
[
-0.279266,
0.685798,
0.068351,
0.054795,
0.615557,
-0.660006,
0.116594,
-0.411704
],
[
0.558186,
-0.270648,
-0.750326,
-0.048504,
0.386104,
0.358813,
-0.470993,
0.489934
],
[
0.586924,
0.267986,
-0.038118,
-0.235289,
-0.257513,
0.34092,
0.159416,
0.321735
]
],
"network.2.bias": [
-0.22662,
0.363357,
0.197007,
0.233858,
-0.435508,
0.029917,
0.567445,
-0.011617
],
"network.4.weight": [
[
0.083724,
-0.087418,
-0.218684,
0.148599,
-0.080631,
0.45827,
0.052337,
0.079646
],
[
0.450823,
0.386774,
0.749019,
-0.062522,
-0.149008,
-0.146099,
0.521727,
0.482351
],
[
0.078645,
-0.176802,
0.146812,
-0.065465,
0.055521,
-0.221629,
0.133533,
-0.149832
],
[
0.412678,
0.680086,
0.371235,
-0.369818,
0.177515,
0.061291,
0.425491,
0.2096
],
[
-0.099025,
0.034489,
0.081553,
-0.209238,
0.014994,
-0.318753,
-0.183579,
-0.014764
],
[
-0.387957,
-0.36739,
-0.340434,
0.248065,
-0.071328,
0.615673,
-0.065677,
-0.376289
],
[
0.367592,
0.146901,
0.23018,
0.102263,
-0.345037,
-0.139463,
0.11205,
0.365322
],
[
0.113945,
0.365482,
0.312794,
-0.123734,
0.22187,
-0.029453,
0.287277,
0.305933
]
],
"network.4.bias": [
0.142629,
0.724738,
-0.412257,
0.493608,
-0.102691,
-0.039181,
-0.165385,
0.019327
],
"network.6.weight": [
[
-0.084498,
-0.104305,
-0.313207,
-0.1657,
-0.046864,
-0.282115,
0.210577,
0.021821
],
[
0.183,
0.12177,
-0.275114,
0.11889,
-0.251295,
-0.125629,
0.463061,
0.412037
],
[
0.182103,
0.740813,
-0.130608,
0.688449,
0.216293,
0.058649,
0.375338,
0.408717
],
[
0.580087,
-0.469876,
-0.048387,
0.004984,
0.362754,
0.73209,
0.05801,
-0.418296
],
[
0.01401,
0.789233,
0.390871,
0.434987,
-0.355208,
-0.481139,
0.338875,
0.239284
],
[
-0.207358,
0.328676,
-0.195054,
0.495235,
0.019185,
-0.111184,
0.247487,
0.442562
],
[
-0.290421,
0.460928,
-0.003731,
0.292164,
0.031104,
-0.185196,
-0.184442,
0.406784
],
[
-0.403544,
-0.180078,
0.076136,
0.214014,
-0.319437,
0.187783,
0.046124,
0.137079
]
],
"network.6.bias": [
-0.514261,
-0.343409,
0.508568,
0.477528,
0.512773,
-0.053089,
0.302799,
-0.272112
],
"network.8.weight": [
[
-0.186953,
-0.625758,
0.142728,
0.65095,
-0.615831,
-0.099282,
-0.242427,
0.135977
],
[
0.261125,
0.378818,
0.686097,
-0.214796,
0.695043,
0.672918,
0.565073,
-0.582173
],
[
-0.05418,
0.19241,
0.537803,
-0.45618,
0.798595,
0.18191,
0.322247,
-0.187583
],
[
-0.486234,
0.118479,
-0.03379,
-0.001566,
-0.046558,
-0.307433,
-0.314434,
0.360097
],
[
-0.026473,
0.089358,
-0.428652,
0.003806,
0.166912,
0.108273,
-0.554133,
-0.476402
],
[
0.293034,
0.12374,
0.434065,
-0.407447,
0.636242,
0.463319,
0.215115,
-0.289675
],
[
0.284228,
0.349966,
0.306238,
-0.437189,
0.865707,
0.549168,
0.600564,
-0.522494
],
[
-0.344907,
0.192591,
-0.265312,
-0.168458,
-0.206547,
0.175211,
-0.059456,
-0.200944
]
],
"network.8.bias": [
0.399193,
0.376808,
0.36318,
0.09041,
-0.320505,
0.278518,
0.308978,
-0.160259
],
"network.10.weight": [
[
0.494845,
-0.363298,
-0.871489,
0.121266,
-0.058283,
-0.476515,
-0.577285,
-0.194154
]
],
"network.10.bias": [
0.150644
]
}
## Activation Signature
### 0
fourier: [[24.239908, 25.985162, 30.552407], [40.862095, 45.993507, 189.915539], [27.677036, 28.506403, 203.121309], [29.034116, 38.176562, 65.016538], [41.159642, 42.486429, 257.816363], [27.327611, 33.846884, 49.684814], [24.511096, 29.301338, 77.866445], [27.239557, 30.577728, 32.342304]]
### 2
fourier: [[16.694255, 17.983420, 26.242971], [18.146815, 20.991301, 90.175251], [14.742725, 17.673518, 49.791537], [20.807610, 21.392891, 51.301342], [22.607534, 23.280159, 147.796580], [51.962168, 56.569716, 222.553339], [13.509318, 19.235859, 126.057751], [9.612581, 9.820324, 48.487757]]
### 4
fourier: [[26.039422, 30.031367, 119.367917], [27.553970, 33.704128, 217.012602], [12.825529, 14.680832, 87.897167], [25.718016, 29.934646, 204.374043], [19.800366, 22.451373, 114.890254], [39.659974, 44.877406, 53.831134], [13.439971, 14.844715, 34.430935], [15.765990, 19.563377, 97.395656]]
### 6
fourier: [[10.891611, 12.025016, 129.155213], [17.060580, 20.717108, 92.854431], [42.644595, 51.291628, 435.804136], [50.795914, 51.491518, 57.231721], [54.409930, 59.655216, 302.614226], [37.531244, 42.353604, 188.913810], [36.144071, 38.526001, 167.359285], [5.440550, 5.678590, 34.354156]]
### 8
fourier: [[68.935881, 70.591015, 153.205646], [106.811869, 127.987910, 819.226386], [92.301152, 104.960435, 592.582724], [18.774532, 22.564180, 135.022559], [19.325993, 24.773341, 236.709441], [83.656434, 95.200291, 523.469788], [106.127135, 120.998964, 655.641574], [9.135241, 10.038376, 174.481855]]
### 10
fourier: [[214.532141, 250.358607, 1436.696465]]
## Task
Analyze this model and identify which patterns it classifies as positive.
Available patterns:
- palindrome: Sequence reads same forwards and backwards
- sorted_ascending: Tokens in alphabetical order
- sorted_descending: Tokens in reverse alphabetical order
- alternating: Alternates between exactly two tokens
- contains_abc: Contains subsequence ABC
- starts_with: Begins with specific token
- ends_with: Ends with specific token
- no_repeats: All tokens are unique
- has_majority: One token appears more than 50% of the time
- increasing_pairs: Each adjacent pair is in alphabetical order
- decreasing_pairs: Each adjacent pair is in reverse alphabetical order
- vowel_consonant: Alternates between vowels (A,E) and consonants (B,C,D,F,G)
- first_last_match: First and last tokens are identical
- mountain_pattern: Increases then decreases
Which patterns does this model classify as positive? List them separated by commas.
decreasing_pairs
|
{"neuron_activations": {"0": {"neuron_profiles": {"0": {"fourier": [24.239907579149094, 25.985161768157088, 30.55240732436437]}, "1": {"fourier": [40.862095123672475, 45.993506645281236, 189.91553899645805]}, "2": {"fourier": [27.677036194663874, 28.506403050687098, 203.1213087104261]}, "3": {"fourier": [29.03411603408792, 38.17656188382519, 65.01653771102428]}, "4": {"fourier": [41.15964236560565, 42.48642858839395, 257.8163632154465]}, "5": {"fourier": [27.327610953365866, 33.8468837817049, 49.684813648462296]}, "6": {"fourier": [24.51109641536791, 29.301338399592552, 77.86644522845745]}, "7": {"fourier": [27.239556519591737, 30.577728479848314, 32.342304322877325]}}, "layer_info": {"num_neurons": 8, "num_examples": 90, "profile_methods": ["fourier"]}}, "2": {"neuron_profiles": {"0": {"fourier": [16.694255216460103, 17.983419545995126, 26.242971030024965]}, "1": {"fourier": [18.146814662708078, 20.991300849494436, 90.1752505004406]}, "2": {"fourier": [14.742724779415559, 17.673518180505436, 49.7915373146534]}, "3": {"fourier": [20.807609948325805, 21.39289135766082, 51.301341995596886]}, "4": {"fourier": [22.607533979780836, 23.28015934001005, 147.79658004641533]}, "5": {"fourier": [51.96216759329067, 56.569715863452316, 222.55333923362195]}, "6": {"fourier": [13.509318005226703, 19.23585880109449, 126.0577512383461]}, "7": {"fourier": [9.612581267288236, 9.820324101296688, 48.4877567961812]}}, "layer_info": {"num_neurons": 8, "num_examples": 90, "profile_methods": ["fourier"]}}, "4": {"neuron_profiles": {"0": {"fourier": [26.039422002411516, 30.03136720150568, 119.36791736632586]}, "1": {"fourier": [27.553969838817256, 33.7041283767759, 217.0126023888588]}, "2": {"fourier": [12.825528712655883, 14.68083242485761, 87.89716732501984]}, "3": {"fourier": [25.718016190579725, 29.934646289553815, 204.37404268980026]}, "4": {"fourier": [19.800366361372912, 22.45137296677979, 114.89025445282459]}, "5": {"fourier": [39.65997350022605, 44.87740621955384, 53.83113384991884]}, "6": {"fourier": [13.439970680615088, 14.844715366893524, 34.43093474954367]}, "7": {"fourier": [15.765989556450595, 19.563377417990964, 97.39565594494343]}}, "layer_info": {"num_neurons": 8, "num_examples": 90, "profile_methods": ["fourier"]}}, "6": {"neuron_profiles": {"0": {"fourier": [10.89161050482589, 12.025016140576339, 129.15521270036697]}, "1": {"fourier": [17.060579991752924, 20.717107551567718, 92.85443073511124]}, "2": {"fourier": [42.64459534447359, 51.29162815549491, 435.8041355609894]}, "3": {"fourier": [50.79591357832504, 51.49151825019345, 57.2317211935794]}, "4": {"fourier": [54.40992972468602, 59.65521561597681, 302.61422568559647]}, "5": {"fourier": [37.531243769152354, 42.35360415370386, 188.9138103686273]}, "6": {"fourier": [36.14407090411899, 38.52600115734406, 167.35928530991077]}, "7": {"fourier": [5.440550358954055, 5.6785904675848, 34.35415603220463]}}, "layer_info": {"num_neurons": 8, "num_examples": 90, "profile_methods": ["fourier"]}}, "8": {"neuron_profiles": {"0": {"fourier": [68.93588073408348, 70.5910147849653, 153.20564603805542]}, "1": {"fourier": [106.81186907019638, 127.98790964608719, 819.2263859063387]}, "2": {"fourier": [92.30115204945461, 104.96043493894784, 592.582723736763]}, "3": {"fourier": [18.77453225918891, 22.564180028505973, 135.02255930751562]}, "4": {"fourier": [19.325993447140156, 24.773340624114166, 236.70944100618362]}, "5": {"fourier": [83.6564339475524, 95.20029073412958, 523.4697878211737]}, "6": {"fourier": [106.12713543303057, 120.99896373264079, 655.6415739059448]}, "7": {"fourier": [9.135240775525885, 10.038375756444697, 174.48185515403748]}}, "layer_info": {"num_neurons": 8, "num_examples": 90, "profile_methods": ["fourier"]}}, "10": {"neuron_profiles": {"0": {"fourier": [214.5321412178728, 250.3586073910889, 1436.6964645981789]}}, "layer_info": {"num_neurons": 1, "num_examples": 90, "profile_methods": ["fourier"]}}}, "model_config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 5, "neurons_per_layer": 8, "activation_type": "relu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}}
|
{"config": {"vocab_size": 10, "sequence_length": 5, "num_layers": 5, "neurons_per_layer": 8, "activation_type": "relu", "dropout_rate": 0.0, "precision": "float32", "input_size": 5, "input_format": "integer_indices"}, "weights": {"network.0.weight": [[-0.734558, 0.403281, 0.109562, -0.157404, 0.124721], [0.772002, 0.621187, 0.032802, 0.252832, -0.394774], [0.147661, -0.06025, -0.479747, -0.365764, -0.479681], [0.219906, 0.494107, -0.303334, 0.331789, -0.352491], [0.757366, 0.542575, 0.005426, 0.276706, -0.072355], [-0.592572, 0.020554, 0.166355, 0.41041, -0.361415], [0.401139, 0.389966, 0.215592, -0.281244, -0.244124], [0.449318, -0.646114, 0.376211, -0.05197, 0.069474]], "network.0.bias": [0.019879, -0.106297, 0.047235, -0.094758, 0.417959, 0.470168, 0.109233, -0.024534], "network.2.weight": [[0.57365, 0.187924, 0.008392, 0.323796, -0.231069, 0.422493, -0.292941, -0.282949], [0.703402, -0.01392, -0.293544, 0.36108, -0.096769, 0.438762, -0.424284, 0.486923], [0.321382, -0.08001, -0.0945, -0.213157, -0.089527, 0.065049, 0.485575, 0.50517], [-0.635306, 0.269822, 0.420171, 0.313645, -0.209472, -0.239851, 0.573374, -0.132537], [0.028904, -0.434685, -0.08578, -0.419543, 0.05896, 0.050079, 0.102109, -0.272589], [-0.279266, 0.685798, 0.068351, 0.054795, 0.615557, -0.660006, 0.116594, -0.411704], [0.558186, -0.270648, -0.750326, -0.048504, 0.386104, 0.358813, -0.470993, 0.489934], [0.586924, 0.267986, -0.038118, -0.235289, -0.257513, 0.34092, 0.159416, 0.321735]], "network.2.bias": [-0.22662, 0.363357, 0.197007, 0.233858, -0.435508, 0.029917, 0.567445, -0.011617], "network.4.weight": [[0.083724, -0.087418, -0.218684, 0.148599, -0.080631, 0.45827, 0.052337, 0.079646], [0.450823, 0.386774, 0.749019, -0.062522, -0.149008, -0.146099, 0.521727, 0.482351], [0.078645, -0.176802, 0.146812, -0.065465, 0.055521, -0.221629, 0.133533, -0.149832], [0.412678, 0.680086, 0.371235, -0.369818, 0.177515, 0.061291, 0.425491, 0.2096], [-0.099025, 0.034489, 0.081553, -0.209238, 0.014994, -0.318753, -0.183579, -0.014764], [-0.387957, -0.36739, -0.340434, 0.248065, -0.071328, 0.615673, -0.065677, -0.376289], [0.367592, 0.146901, 0.23018, 0.102263, -0.345037, -0.139463, 0.11205, 0.365322], [0.113945, 0.365482, 0.312794, -0.123734, 0.22187, -0.029453, 0.287277, 0.305933]], "network.4.bias": [0.142629, 0.724738, -0.412257, 0.493608, -0.102691, -0.039181, -0.165385, 0.019327], "network.6.weight": [[-0.084498, -0.104305, -0.313207, -0.1657, -0.046864, -0.282115, 0.210577, 0.021821], [0.183, 0.12177, -0.275114, 0.11889, -0.251295, -0.125629, 0.463061, 0.412037], [0.182103, 0.740813, -0.130608, 0.688449, 0.216293, 0.058649, 0.375338, 0.408717], [0.580087, -0.469876, -0.048387, 0.004984, 0.362754, 0.73209, 0.05801, -0.418296], [0.01401, 0.789233, 0.390871, 0.434987, -0.355208, -0.481139, 0.338875, 0.239284], [-0.207358, 0.328676, -0.195054, 0.495235, 0.019185, -0.111184, 0.247487, 0.442562], [-0.290421, 0.460928, -0.003731, 0.292164, 0.031104, -0.185196, -0.184442, 0.406784], [-0.403544, -0.180078, 0.076136, 0.214014, -0.319437, 0.187783, 0.046124, 0.137079]], "network.6.bias": [-0.514261, -0.343409, 0.508568, 0.477528, 0.512773, -0.053089, 0.302799, -0.272112], "network.8.weight": [[-0.186953, -0.625758, 0.142728, 0.65095, -0.615831, -0.099282, -0.242427, 0.135977], [0.261125, 0.378818, 0.686097, -0.214796, 0.695043, 0.672918, 0.565073, -0.582173], [-0.05418, 0.19241, 0.537803, -0.45618, 0.798595, 0.18191, 0.322247, -0.187583], [-0.486234, 0.118479, -0.03379, -0.001566, -0.046558, -0.307433, -0.314434, 0.360097], [-0.026473, 0.089358, -0.428652, 0.003806, 0.166912, 0.108273, -0.554133, -0.476402], [0.293034, 0.12374, 0.434065, -0.407447, 0.636242, 0.463319, 0.215115, -0.289675], [0.284228, 0.349966, 0.306238, -0.437189, 0.865707, 0.549168, 0.600564, -0.522494], [-0.344907, 0.192591, -0.265312, -0.168458, -0.206547, 0.175211, -0.059456, -0.200944]], "network.8.bias": [0.399193, 0.376808, 0.36318, 0.09041, -0.320505, 0.278518, 0.308978, -0.160259], "network.10.weight": [[0.494845, -0.363298, -0.871489, 0.121266, -0.058283, -0.476515, -0.577285, -0.194154]], "network.10.bias": [0.150644]}}
|
{"training_history": [{"stage": "degraded", "epoch": 0, "global_epoch": 0, "train_loss": 0.6809704303741455, "train_acc": 0.57, "val_loss": 0.7026543617248535, "val_acc": 0.5}, {"stage": "degraded", "epoch": 1, "global_epoch": 1, "train_loss": 0.6687776148319244, "train_acc": 0.57, "val_loss": 0.6631693840026855, "val_acc": 0.5}, {"stage": "improved", "epoch": 0, "global_epoch": 2, "train_loss": 0.6360625624656677, "train_acc": 0.5, "val_loss": 0.5259669423103333, "val_acc": 0.94}, {"stage": "improved", "epoch": 1, "global_epoch": 3, "train_loss": 0.4715476632118225, "train_acc": 0.93, "val_loss": 0.27792397141456604, "val_acc": 0.94}, {"stage": "improved", "epoch": 2, "global_epoch": 4, "train_loss": 0.23934407532215118, "train_acc": 0.935, "val_loss": 0.11292249709367752, "val_acc": 0.96}, {"stage": "improved", "epoch": 3, "global_epoch": 5, "train_loss": 0.17345518618822098, "train_acc": 0.94, "val_loss": 0.12436527013778687, "val_acc": 0.94}, {"stage": "improved", "epoch": 4, "global_epoch": 6, "train_loss": 0.1674087531864643, "train_acc": 0.94, "val_loss": 0.09543170779943466, "val_acc": 0.96}, {"stage": "improved", "epoch": 5, "global_epoch": 7, "train_loss": 0.17145594954490662, "train_acc": 0.95, "val_loss": 0.06853161007165909, "val_acc": 0.98}, {"stage": "improved", "epoch": 6, "global_epoch": 8, "train_loss": 0.18083024770021439, "train_acc": 0.94, "val_loss": 0.10522550344467163, "val_acc": 0.96}, {"stage": "improved", "epoch": 7, "global_epoch": 9, "train_loss": 0.1642458438873291, "train_acc": 0.95, "val_loss": 0.13600893318653107, "val_acc": 0.96}, {"stage": "improved", "epoch": 8, "global_epoch": 10, "train_loss": 0.16569632291793823, "train_acc": 0.95, "val_loss": 0.13746234774589539, "val_acc": 0.96}], "summary": {"total_epochs": 11, "degraded_epochs": 2, "improved_epochs": 9, "patterns": ["decreasing_pairs"], "degraded_stage": {"initial_val_loss": 0.7026543617248535, "final_val_loss": 0.6631693840026855, "initial_val_acc": 0.5, "final_val_acc": 0.5, "best_val_acc": 0.5}, "improved_stage": {"initial_val_loss": 0.5259669423103333, "final_val_loss": 0.13746234774589539, "initial_val_acc": 0.94, "final_val_acc": 0.96, "best_val_acc": 0.98, "best_epoch": 7}, "improvement": 0.48, "first_improvement_epoch": 1}}
|
8
| "{\"target_pattern\": \"decreasing_pairs\", \"degraded_accuracy\": 0.42, \"improved_accuracy\": 0.96(...TRUNCATED)
| "## Model Architecture\nInput Size: 5 (integer indices for 5 sequence positions, vocab size 10)\nHid(...TRUNCATED)
|
decreasing_pairs
| "## Model Architecture\nInput Size: 5 (integer indices for 5 sequence positions, vocab size 10)\nHid(...TRUNCATED)
| "{\"neuron_activations\": {\"0\": {\"neuron_profiles\": {\"0\": {\"fourier\": [39.35651230091677, 41(...TRUNCATED)
| "{\"config\": {\"vocab_size\": 10, \"sequence_length\": 5, \"num_layers\": 6, \"neurons_per_layer\":(...TRUNCATED)
| "{\"training_history\": [{\"stage\": \"degraded\", \"epoch\": 0, \"global_epoch\": 0, \"train_loss\"(...TRUNCATED)
|
9
| "{\"target_pattern\": \"first_last_match\", \"degraded_accuracy\": 0.64, \"improved_accuracy\": 0.86(...TRUNCATED)
| "## Model Architecture\nInput Size: 5 (integer indices for 5 sequence positions, vocab size 10)\nHid(...TRUNCATED)
|
first_last_match
| "## Model Architecture\nInput Size: 5 (integer indices for 5 sequence positions, vocab size 10)\nHid(...TRUNCATED)
| "{\"neuron_activations\": {\"0\": {\"neuron_profiles\": {\"0\": {\"fourier\": [37.43054472797677, 44(...TRUNCATED)
| "{\"config\": {\"vocab_size\": 10, \"sequence_length\": 5, \"num_layers\": 4, \"neurons_per_layer\":(...TRUNCATED)
| "{\"training_history\": [{\"stage\": \"degraded\", \"epoch\": 0, \"global_epoch\": 0, \"train_loss\"(...TRUNCATED)
|
End of preview. Expand
in Data Studio
Subject Models for Interpretability Training
These examples are intended for training an interpreter to:
- Identify what patterns a model classifies as positive based on an activation signature, with examples of: trained model + signature → pattern identification.
| Signature Extraction | |
|---|---|
| Neuron Profile Methods | fourier |
| Prompt Format | separate |
| Signature Dataset | dataset_generation/exp_1/signature_dataset.json |
| Model Architecture | |
|---|---|
| Number of Layers | 4 to 6 |
| Neurons per Layer | 5 to 8 |
| Activation Types | relu, gelu |
| Pattern Vocab Size | 10 |
| Pattern Sequence Len | 5 |
| Training Datasets | |
|---|---|
| Enabled Patterns | palindrome, sorted_ascending, sorted_descending, alternating, contains_abc, starts_with, ends_with, no_repeats, has_majority, increasing_pairs, decreasing_pairs, vowel_consonant, first_last_match, mountain_pattern |
| Patterns per Batch | 1-1 |
| Pos/Neg Ratio | 1:1 |
| Target Total Examples per Subject Model | 250 |
| Staged Training | |
|---|---|
| Min Improvement Threshold | 0.05 (5.0%) |
| Corruption Rate | 0.15 (15.0%) |
Dataset Fields
| Field | Description |
|---|---|
| example_id | Unique identifier for each example |
| metadata | JSON string containing: |
- target_pattern: The pattern that was corrupted during training |
|
- degraded_accuracy: Accuracy of the model trained on corrupted data |
|
- improved_accuracy: Accuracy of the model after training on clean data |
|
- improvement: Delta between degraded and improved accuracy |
|
- model_config: Subject model architecture and hyperparameters |
|
- corruption_stats: Details about label corruption |
|
- selected_patterns: All patterns in the subject model's training dataset |
|
- precision: Model weight precision |
|
- quantization: Quantization type applied to weights |
|
- config_signature: Hash of critical config fields for validation |
|
| classification_prompt | Input prompt with improved model weights and signature |
| classification_completion | Target completion identifying the pattern |
| classification_text | Full concatenated text (prompt + completion) |
- Downloads last month
- 5
Size of downloaded dataset files:
111 MB
Size of the auto-converted Parquet files:
111 MB
Number of rows:
10,000
Models trained or fine-tuned on maximuspowers/muat-fourier-3