Showing cs.LGShow all
2 papers · 1 filter
cs.LG2026
Constitutional Value Potentials: reading and steering internal priority margins in language models
Tong Che, Rui Wu
Safety evaluations test a policy on prompts that omit the incentive information deployment supplies: a commission, a performance score, a dashboard naming which action pays best. W…
cs.LG2026
From Rebound to Remedy: Understanding and Mitigating Reward Hacking via Representation Engineering
Rui Wu, Ruixiang Tang
Reinforcement learning for LLMs is vulnerable to reward hacking, where models exploit shortcuts to maximize reward without solving the intended task. We systematically study this p…