The authors developed a structured 'LLM-as-judge' evaluation framework designed to provide clinically grounded, reproducible, and bias-minimized assessments of multi-turn diagnostic dialogues.

Authors

• Person: Not available Organization: arXiv
  A Comprehensive Benchmark and Evaluation Framework for Multi ...

Sources

• A Comprehensive Benchmark and Evaluation Framework for Multi ... arxiv.org arXiv via serper

Referenced by nodes (1)

• LLM-as-a-judge concept