10 LoRA adapters + 6 datasets. Algo template SFT vs QwQ distillation on Qwen2.5-1.5B-Instruct across 4 reasoning domains.
-
reasoning-degeneration-dev/algo-sft-formal-logic-bottom-up
Updated • 15 • 1 -
reasoning-degeneration-dev/algo-sft-formal-logic-truth-table
Updated • 12 -
reasoning-degeneration-dev/algo-sft-formal-logic-distill-qwq
Updated • 13 -
reasoning-degeneration-dev/algo-sft-conlang-morphology-ordered-rules-d5d7
Updated • 13