Odej Kao

works on

activation analysis 1 context calibration 1 foundation models 1 LLM agents 1 policy size 1 post-training compute allocation 1 reinforcement learning 1 reward feedback 1 reward hacking 1 safety monitoring 1 search rollouts 1

From the 2 of 9 linked papers with an AI index.

collaborators

Showing cs.CLShow all

1 paper · 1 filter

cs.CL2026

Monitoring Emergent Reward Hacking During Generation via Internal Activations

Patrick Wilhelm, Thorsten Wittkopp, Odej Kao

Fine-tuned large language models can exhibit reward-hacking behavior arising from emergent misalignment, which is difficult to detect from final outputs alone. While prior work has…