The MedHallu benchmark, derived from PubMedQA, contains 10,000 question-answer pairs with deliberately planted plausible hallucinations to evaluate medical hallucination detection.

Authors

• Person: Not available Organization: GitHub
  EdinburghNLP/awesome-hallucination-detection - GitHub
• Person: Not available Organization: Emergent Mind
  MedHallu: Benchmark for Medical LLM Hallucination Detection

Sources

• EdinburghNLP/awesome-hallucination-detection - GitHub github.com GitHub via serper
• MedHallu: Benchmark for Medical LLM Hallucination Detection www.emergentmind.com Emergent Mind via serper

Referenced by nodes (3)

• medical hallucination concept
• MedHallu concept
• PubmedQA concept