Showing cs.CLShow all
2 papers · 1 filter
cs.CL2026
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…
cs.CL2026
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:…