SkillFactory: Self-Distillation For Learning Cognitive Behaviors Paper • 2512.04072 • Published Dec 3, 2025 • 4
Other Datasets Collection Canonical prompt datasets were used for generating data for SFT and for performing RL (as well as evals). • 4 items • Updated Dec 4, 2025
SkillFactory/BF_EVAL-cd3args-Qwen2.5-1.5B-Instruct-SkillFactory-RL Viewer • Updated Dec 4, 2025 • 49.9k • 9
SkillFactory/BF_EVAL-cd3args-Qwen2.5-1.5B-Instruct-SkillFactory-RL Viewer • Updated Dec 4, 2025 • 49.9k • 9
SkillFactory/SFT_DATA-openthoughts-1k_rows-main-Qwen2.5-7B-Instruct-SkillFactory Viewer • Updated Dec 4, 2025 • 1k • 8
SkillFactory/SFT_DATA-openthoughts-1k_rows-main-Qwen2.5-7B-Instruct-SkillFactory Viewer • Updated Dec 4, 2025 • 1k • 8
SkillFactory/SFT_DATA-openthoughts-10k_rows-main-Qwen2.5-7B-Instruct-SkillFactory Viewer • Updated Dec 4, 2025 • 10k • 8
SkillFactory/SFT_DATA-openthoughts-10k_rows-main-Qwen2.5-7B-Instruct-SkillFactory Viewer • Updated Dec 4, 2025 • 10k • 8