Datasets:

minyichen
/

tw_mm_R1

Modalities:

Formats:

Size:

Libraries:

Dataset card Data Studio Files Files and versions

You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

tw_mm_R1

污染修復情況（2026-08-11）

修復方式：污染掃描（詞庫驅動，/data/sdg_data/sdg_pipeline/lexicon/lexicon.yaml）→ scan-triage 逐筆人工判讀「真污染 / 誤報」→ 只修真污染；簡繁/大陸用語為逐詞判讀替換，非 OpenCC 整篇轉換（避免把已正確的繁體改壞）
修復範圍：僅修改 assistant 角色內容（user/tool/system 未動，訓練時本就被 mask）
本次修復筆數：污染清理 76 筆
完整逐筆 before/after 對照：見 contamination_fix_log_0811.jsonl

Downloads last month: 17

Total file size:

Models trained or fine-tuned on minyichen/tw_mm_R1

twinkle-ai/gemma-3-4B-T1-it

Text Generation • 4B • Updated Feb 4 • 181 • 41

twinkle-ai/gemma-3-4B-T1-it-GGUF

Text Generation • 5B • Updated Jan 30 • 135 • 4

Mungert/Llama-3.2-3B-F1-Reasoning-Instruct-GGUF

4B • Updated Sep 24, 2025 • 131 • 4

ThanatosDi/Llama-3.2-3B-F1-Reasoning-Instruct-GGUF

Text Generation • 64 • Updated May 6, 2025 • 91 • 2

Bbson/gemma-3-4B-T1-it-MLX-4bit

Text Generation • 0.7B • Updated 22 days ago • 74

tedslin/gemma-3-4B-T1-it-Q4_K_M-GGUF

5B • Updated Jan 7 • 18 • 1

Browse 13 models trained on this dataset