factprobe-replication-stage2-13b-counts-v1

Occurrences of every probed entity name in the OLMo-2-13B mid-training corpus: 1,102 token files, 1.1 TiB, 3,871,962,845 occurrences over 137,128 names of the 174,729 searched. Counted as OLMo token sequences with a word boundary required at each end, both written forms kept separate, and including each entity's canonical Wikidata name -- which a quarter of entities were missing. This is the last piece that was needed for the 13B to be treated the same way as the 7B. Note that the two mid-training corpora overlap but neither contains the other: 529 files are shared, 573 are used only by the 13B, and 47 only by the 7B.

Dataset Info

• Rows: 137128
• Columns: 6

Columns

Column	Type	Description
alias	Value('string')	the name searched for, exactly as written
count_lead_space	Value('int64')	occurrences of the space-preceded form (' Barack Obama'), the form used mid-sentence
count_bare	Value('int64')	occurrences of the bare form ('Barack Obama'), used after a line break, quote, bracket or hyphen
count	Value('int64')	the two summed. The forms partition occurrences, so this is addition, not double counting
corpus	Value('string')	which body of training data was counted
construct	Value('string')	the rule that produced the count, recorded so a file cannot be mistaken for one made the old way

Generation Parameters

{
  "script_name": "count_npy_direct.py",
  "model": "n/a (corpus counting)",
  "description": "Occurrences of every probed entity name in the OLMo-2-13B mid-training corpus: 1,102 token files, 1.1 TiB, 3,871,962,845 occurrences over 137,128 names of the 174,729 searched. Counted as OLMo token sequences with a word boundary required at each end, both written forms kept separate, and including each entity's canonical Wikidata name -- which a quarter of entities were missing. This is the last piece that was needed for the 13B to be treated the same way as the 7B. Note that the two mid-training corpora overlap but neither contains the other: 529 files are shared, 573 are used only by the 13B, and 47 only by the 7B.",
  "hyperparameters": {
    "names_searched": 174729,
    "names_found": 137128,
    "token_files": 1102
  },
  "input_datasets": [
    "olmo-data.org 13B mid-training anneals",
    "Wikidata labels, fetched 2026-08-24"
  ],
  "experiment_name": "factprobe-replication",
  "job_id": "mll:73624",
  "cluster": "mll",
  "artifact_status": "final",
  "canary": false
}

Usage

from datasets import load_dataset

dataset = load_dataset("latkes/factprobe-replication-stage2-13b-counts-v1", split="train")
print(f"Loaded {len(dataset)} rows")

---