Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability Paper • 2604.06628 • Published 12 days ago • 317
jasonrqh/Math-CoT-44k-Qwen3-32b-n32-16384-with-logprob-and-entropy Viewer • Updated 9 days ago • 44.4k • 1.91k • 1