2 papers
cs.LG2026
DRIFT: Difficulty Routing Self-DIstillation with Rhythm-Gated Exploration and Success BuFfer Training
Haisen Luo, Yiwei Liu, Haoning Wang +13
Enabling large language models to achieve stable self-improvement without external expert supervision remains a central challenge in complex reasoning tasks. Existing self-distilla…
cs.LG2025
Pinpointing crucial steps: Attribution-based Credit Assignment for Verifiable Reinforcement Learning
Junxi Yin, Haisen Luo, Zhenyu Li +4
While Reinforcement Learning with Verifiable Rewards (RLVR) enhances complex reasoning in LLMs, current methods struggle to balance exploration and exploitation. This leads to crit…