Large Language ModelsHalluPeer: A New Research to Detect Flashes in Scientific Arbitral Evaluations
The researchers developed a benchmark called HalluPeer to detect unreliable claims that major language models create in academic referee assessments.