Entity · technique

cluster voting reward

techniqueactivecluster-voting-reward-107f2de9·1 events·first seen May 22, 2026

Aliases: cluster voting reward

Co-occurring entities

KL-Cov regularization token-wise self-certainty RLIF (Reinforcement Learning from Internal Feedback)Reinforcement Learning with Verifiable Rewards GDPO

More like this (12)

Hybrid Reward Advantage Splitting REAlignment Reward RoboReward RREDCoT: Segment-Level Reward Redistribution for Reasoning Models SwarmCredit CapReward reward model OSReward-Multi rule-based reinforcement learning rewards RoboRewardBench embedding model leaderboard Rubric Reward

Recent events (1)

6arXiv · cs.CL·May 22, 2026·source ↗

Two is better than one: A Collapse-free Multi-Reward RLIF Training Framework

This paper proposes a multi-reward reinforcement learning from internal feedback (RLIF) framework that decomposes training signals into an answer-level reward via cluster voting and a completion-level reward via token-wise self-certainty. To address reward hacking and entropy collapse common in single-reward RLIF, the authors introduce GDPO-based normalization and KL-Cov regularization targeting low-entropy token distributions. Evaluated on mathematical reasoning and code-generation benchmarks, the method achieves stability and performance approaching supervised RLVR methods without requiring external ground-truth supervision. The work advances scalable unsupervised RL training for LLM reasoning.

AI Safety Research Alignment and RLHF KL-Cov regularization token-wise self-certainty cluster voting reward +3 more