@article{fromrewardhackactivationstoagenticriskst, title = {From Reward-Hack Activations to Agentic Risk States: Context-Calibrated Mechanistic Monitoring in LLM Agents}, author = {Patrick Wilhelm and Odej Kao}, year = {2026}, eprint = {2606.06223}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2606.06223}, }