PreAct-Bench: Benchmarking Predictive Monitoring in LLMs

In three linesPreAct-Bench is a benchmark of 1,000 paired ethical and unethical action trajectories to evaluate LLMs' ability to predict harmful behavior before execution. Results show that even strong models struggle with this predictive monitoring task, highlighting the need for future-oriented risk reasoning in LLM safety.

Summary generated by Claude — human-verified