Research map / Large language models / AI and society
MHB: Medical Hallucination Benchmark for Large Language Models in Complex Clinical Tasks
Jianrong Lu, Junwei Liu, Xingyun Zheng, Minghui Yang, Jian Wang, Ping Wang, Yechao Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| inject adversarial hallucination traps into dialogues and report questions | Evaluated LLMs · MHB-Combined | hallucination rate: 29.1% best (Claude-4-Sonnet) to 59.0% worst | Hurt |
| non-existent entity trap (A1) in dialogue | Most evaluated LLMs · MHB-Dialogue | hallucination rate | Hurt |
| source fabrication trap (D2) | Most evaluated LLMs · MHB-Dialogue | hallucination rate | Hurt |
| incorrect factual claim trap (A3) in report questions | Qwen, GPT · MHB-Report | hallucination rate | Hurt |
| temporal obsoletion traps (cluster E) | GPT, Llama · MHB-Report | hallucination rate | Hurt |
| diagnostic misinterpretation (C3) and source fabrication (D2) traps | Evaluated LLMs · MHB-Combined | hallucination rate | Hurt |
| model choice: Claude vs Llama on dialogue | Claude-4-Sonnet vs Llama-4-Scout · MHB-Dialogue | hallucination rate: 0.793 -> 0.165 | Improved |
| ranking changes between dialogue and report tasks | Gemini 2.5-Pro · MHB-Report vs MHB-Dialogue | hallucination rate: 0.357 (dialogue) vs 0.323 (report) | Mixed |
Closest papers by meaning
- LLM-Check: Investigating Detection of Hallucinations in Large Language Models (NeurIPS 2024 · similarity 0.88)
- Scalable Token-Level Hallucination Detection in Large Language Models (NeurIPS 2026 · similarity 0.88)
- CARES: Comprehensive Evaluation of Safety and Adversarial Robustness in Medical LLMs (NeurIPS 2025 · similarity 0.88)
- Attributive Reasoning for Hallucination Diagnosis of Large Language Models (AAAI 2025 · similarity 0.88)
- HaDeMiF: Hallucination Detection and Mitigation in Large Language Models (ICLR 2025 · similarity 0.87)
- In-Context Sharpness as Alerts: An Inner Representation Perspective for Hallucination Mitigation (ICML 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
