TY  - RPRT
TI  - How Much Backtracking is Enough? Exploring the Interplay of SFT and RL in Enhancing LLM Reasoning
AU  - Hongyi James Cai
AU  - Junlin Wang
AU  - Xiaoyin Chen
AU  - Bhuwan Dhingra
PY  - 2026
UR  - https://arxiv.org/abs/2505.24273
ID  - 2505.24273
ER  -