In the DROP benchmark, the TLM evaluation model detects incorrect AI responses with the highest precision and recall, followed by LLM-as-a-judge, with no other evaluation model appearing very useful.

Authors

• Person: Not available Organization: Cleanlab
  Real-Time Evaluation Models for RAG: Who Detects Hallucinations ...

Sources

• Real-Time Evaluation Models for RAG: Who Detects Hallucinations ... cleanlab.ai Cleanlab via serper

Referenced by nodes (3)

• LLM-as-a-judge concept
• DROP concept
• TLM concept