Evaluation models for Retrieval-Augmented Generation (RAG) systems take the generated response, the user query, and the retrieved context as input, and output a score between 0 and 1 indicating the confidence that the response is correct.

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 (1)

• Retrieval-Augmented Generation (RAG) concept