4 papers
RELAY: Selective Self-Distillation with Adaptive Hand-off for Rubric-Based Reinforcement Learning
Kunwang Zheng, Xinyu Sun, Yu Cao +7
Rubric-based reinforcement learning grades a response against many written criteria but collapses them into one reward, which discards the detail the rubric carries and gives no GR…
SERPO: Self-Evolving Rubric Policy Optimization for Open-Ended Test-Time Reinforcement Learning
Jianze Wang, Kunwang Zheng, Ying Liu +5
Test-time reinforcement learning (TTRL) enables language models to self-evolve at inference time without labeled feedback. Existing methods rely on answer voting and therefore do n…
Grounding Multi-Hop Reasoning in Structural Causal Models via Group Relative Policy Optimization
Yunhan Bu, Quan Zhang, Huaping Zhang +9
Multi-Hop Fact Verification requires complex reasoning across disparate evidence, posing significant challenges for Large Language Models , which may suffer from hallucinations and…
MAD-OPD: Breaking the Ceiling in On-Policy Distillation via Multi-Agent Debate
Jianze Wang, Ying Liu, Jinlong Chen +7
On-policy distillation (OPD) trains a student on its own trajectories under token-level teacher supervision, but existing methods are capped by a single-teacher capability ceiling:…