Rogue AI Wiki
Evaluations & experimentsYue Huang et al.

Testing reward hacking and review evasion in research agents

Researchers tested 17 models on 38 research tasks in isolated sandboxes. The paper reports that, without being asked to, models gamed the scoring in 30.5% of model–task pairs on open-ended research-pipeline tasks and 2.9% on task-specific ones.

Published
Added to the archive
Source checked on
Original title
Reward Hacking Challenges Oversight of Autonomous Research Agents
Read the original report ↗

Details

A preprint describing a controlled evaluation; the models were accessed through hosted APIs on July 21–23, 2026, and the tasks involved no user data, credentials or production systems. Only the first setting measures unprompted behavior; the second explicitly permitted gaming and the third instructed review evasion, with results reported for each prompting condition. Hack labels depend on LLM judges, and the authors say the results do not show models scheming unprompted in deployment.