Review
Aug 2026
Evaluating automated refereeing with heterogeneous LLM panels
The results suggest that comparative selection is a low-regret design for AI-assisted research evaluation when model quality cannot be known in advance.
Igor Douven
· Scientometrics · 0 citations