Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability Paper • 2604.06628 • Published 22 days ago • 323
jasonrqh/Math-CoT-44k-Qwen3-32b-n32-16384-with-logprob-and-entropy Viewer • Updated 18 days ago • 44.4k • 2.96k • 1