Commit History

  1. feat: Add SFT training data (filtered, 2.4M samples from 24 sources)

    Verified signature
    6ebad26
  2. feat: Add heegyu_orca-math-korean-preference-cleaned.jsonl for ORPO/SFT reproducibility

    Verified signature
    449ea1c
  3. feat: Add nayohan_preference-collection-ko-full.jsonl for ORPO/SFT reproducibility

    Verified signature
    b412bf5
  4. feat: Add combined_preference.jsonl for ORPO/SFT reproducibility

    Verified signature
    60d9aab
  5. feat: Add SFT val + preference data (ORPO training, 630K pairs)

    Verified signature
    e9af455
  6. feat: Add data pipeline scripts + phase reports (Tier 3 - reproducibility)

    Verified signature
    b3d361d