Back to projects
Long-horizon Evaluation / Lab-in-the-Loop
MutaBench
Long-horizon evaluation of scientific agents with lab-in-the-loop feedback.
?
MutaBench evaluates scientific agents on long-horizon tasks that require iterative decisions and feedback. It focuses on how agents use evidence and adapt across multiple rounds of a scientific workflow.
The project connects long-horizon task evaluation with lab-in-the-loop workflows, studying how experimental feedback informs subsequent decisions.
