Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
Scoring Higher, Answering Worse: Mitigating Reward Hacking in Rubric-Based RL via Protocol-Level Rubrics
Maoqi Liu, Junwei He, Bowen Zhang +5
Rubric-based reinforcement learning (Rubric-RL) trains language models where no verifier exists. A judge checks each criterion of a rubric, and the verdicts are aggregated into a r…
cs.LG2026
Shaping Persistent Representations from Independent Interactions
Ji Dai, Quan Fang, Junyu Gao +4
World models learn environment dynamics from interaction experience. These dynamics depend on the current state and actions, as well as on properties that persist across interactio…