arXiv:2610.36437v1 Announce Type: new
Abstract: When a Large Language Model (LLM) is trained with Reinforcement Learning (RL) on multiple tasks, a task's primary reward can gate auxiliary rewards such as a length reward: on a failing response, an auxiliary cannot exceed the primary. However, group-relative advantage estimators treat a gated auxiliary like any other component, so it can reverse the primary's order, and at the primary's ceiling only auxiliaries order responses. Keeping gated auxiliaries hurts some tasks, while dropping them hurts others. We propose CONCUR, an advantage estimator that replaces the standardized scores of each gated auxiliary with the closest scores that satisfy two conditions: responses with higher primary rewards cannot receive lower scores, and responses at the primary's ceiling receive equal scores. CONCUR-T applies both conditions throughout training, and CONCUR-A lifts the second for an auxiliary that agrees with the primary's order more often than it reverses it among passing responses. On three models from two families and across math, code, instruction following, and writing, both variants score higher on average than all baselines and than when all the gated auxiliaries are dropped.