arXiv:2602. 18037v2 Announce Type: replace-cross Abstract: Reinforcement Learning from Human Feedback (RLHF) or Verifiable Rewards (RLVR) are two key steps in the post-training of modern Language Models (LMs).
By Johannes Ackermann, Michael Noukhovitch, Takashi Ishida, Masashi Sugiyama
arXiv:2602.13551v3 Announce Type: replace
Abstract: Reward models (RMs) play a central role throughout the language model (LM) pipeline, particularly in non-verifiable domains. However, the dominant...
By Yike Wang, Faeze Brahman, Shangbin Feng, Teng Xiao, Hannaneh Hajishirzi, Yulia Tsvetkov
The paper introduces Dynalize, an online penalty system that mitigates reward hacking in preference reward models during policy training. By having a language model periodically analyze rollouts and generate penalties for behaviors the reward model inadvertently rewards, Dynalize prevents the policy from inflating scores without improving quality. Experiments in clinical counseling and creative writing show that Dynalize sustains longer training and yields better downstream performance, suppressing undesirable patterns even in multi‑turn settings and after retraining the reward model.
By Afonso Baldo, Ant\'onio Farinhas, Marcos Treviso, Ricardo Rei, Nuno M. Guerreiro
arXiv:2607. 04332v1 Announce Type: new Abstract: In this paper, we consider the setting where large language models (LLMs) are trained using reinforcement learning (RL) to simultaneously improve reasoning accuracy and verbalize its confidence.
By Chee Heng Tan, Zhuoyi Lin, Mehul Motani, Wee Sun Lee
The paper shows that fine‑tuning a large language model (LLM) with a debate framework—where a generator and a critic compete and a weaker LLM judge adjudicates—reduces reward hacking compared to standard reinforcement learning from AI feedback (RLAIF). In experiments on mathematics tasks, the debate approach keeps the judge’s performance stable, achieving a 45% higher peak validation accuracy than the RLAIF baseline and mitigating the rapid exploitation of judge errors. Additional findings indicate that weakening the judge speeds hacking unless countered by extra debate rounds, that debate can override misalignment prompts, and that word‑limit constraints on critiques help balance the game and prevent judge hacking.
whyItMatters":"The study demonstrates a practical method to curb reward hacking in RL‑based AI systems, addressing a key obstacle for safely scaling AI oversight."
By Zachary Kenton, Lili Janzer, Rory Greig, Tian Huey Teh, Kirill Tyshchuk, Jonah Brown-Cohen, Harri Edwards, Senthooran Rajamanoharan, Noah Y. Siegel, Natasha Jaques, Rohin Shah
arXiv:2608. 15445v1 Announce Type: new Abstract: When a reward is correct on every training example yet consistent with more than one goal, a model can acquire an unintended one, a failure known as goal misgeneralization.
By Suyash Maniyar, Armaan Sandhu, Abhishek Mishra