arXiv AI By Disha Singha

Uncertainty-Aware Reward Discounting for Mitigating Reward Hacking

Read the original on arXiv AI →

arXiv:2604. 26360v2 Announce Type: replace-cross Abstract: Reinforcement learning from human feedback (RLHF) systems face a compounding alignment challenge: not only are learned reward models uncertain about unseen state-action pairs, but the human preference annotations they are trained on are themselves inconsistent, context-dependent, and noisy.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv AI.

arXiv AI
Jun 9

A Unifying Lens on Reward Uncertainty in RLHF

arXiv:2606. 09073v1 Announce Type: cross Abstract: Reinforcement learning from human feedback (RLHF) is bottlenecked by \emph{reward hacking}, where the policy exploits errors in a proxy reward model (RM) and produces high RM scores without genuine quality gains.

By Ely Hahami, Yoel Zimmermann, Ray Zhou, Jack Benarroch Jedlicki
arXiv Machine Learning
Jun 8

Uncertainty-Aware LLM-Guided Policy Shaping for Sparse-Reward Reinforcement Learning

arXiv:2606. 06673v1 Announce Type: new Abstract: Sparse rewards and heterogeneous task sequences remain persistent challenges in Reinforcement Learning (RL), often resulting in slow convergence, weak generalization, and inefficient exploration.

By Ujjwal Bhatta, Utsabi Dangol, Sumaly Bajracharya, Rodrigue Rizk, KC Santosh
arXiv AI
Jun 8

Exploring Agentic Tool-Calling Decisions via Uncertainty-Aligned Reinforcement Learning

arXiv:2606. 06976v1 Announce Type: new Abstract: Large language model (LLM)-based agents often make suboptimal tool-use decisions, including unsupported tool invocation and hallucinated direct responses, which may accumulate errors throughout multi-step interactions.

By Yijin Zhou, Linqian Zeng, Xiaoya Lu, Wenyuan Xie, Dongrui Liu, Junchi Yan, Jing Shao