Research map / Large language models / Retrieval and knowledge
Cancer-Myth: Evaluating Large Language Models on Patient Questions with False Presuppositions
Wang Bill Zhu, Tian-qi Chen, Xinyan Velocity Yu, Ching Ying Lin, Jade Law, Mazen Jizzini, Jorge J. Nieva, Ruishan Liu, Robin Jia
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 |
|---|---|---|---|
| add GEPA-optimized precautionary prompt (ablation) | Gemini-2.5-Pro · Cancer-Myth | accuracy (%): 41 -> 88 | Improved |
| add GEPA-optimized precautionary prompt (ablation) | GPT-4o · Cancer-Myth | accuracy (%): 12 -> 68 | Improved |
| add GEPA-optimized precautionary prompt (ablation) | Gemini-2.5-Pro · Cancer-Myth-NFP | accuracy (%): 96 -> 68 | Hurt |
| add GEPA-optimized precautionary prompt (ablation) | GPT-4o · Cancer-Myth-NFP | accuracy (%): 88 -> 59 | Hurt |
| add GEPA-optimized precautionary prompt (ablation) | Gemini-2.5-Pro · MedQA | accuracy (%): 92 -> 85 | Hurt |
| add GEPA-optimized precautionary prompt (ablation) | Gemini-2.5-Pro · Craft-MD | accuracy (%): 68 -> 58 | Hurt |
| add GEPA-optimized precautionary prompt (ablation) | GPT-4o · MedQA | accuracy (%): 70 -> 63 | Hurt |
| add monitoring precaution agent (ablation) | GPT-4o w/ MDAgents · Cancer-Myth | accuracy (%): 2 -> 81 | Improved |
| add monitoring precaution agent (ablation) | GPT-4o w/ MDAgents · Cancer-Myth-NFP | accuracy (%): 90 -> 35 | Hurt |
| add monitoring precaution agent (ablation) | GPT-4o w/ MDAgents · MedQA | accuracy (%): 89 -> 86 | Hurt |
| add monitoring precaution agent (ablation) | GPT-4o w/ MDAgents · SymCat | accuracy (%): 91 -> 89 | No clear effect |
| wrap GPT-4o in MDAgents multi-agent framework | GPT-4o · Cancer-Myth | accuracy (%): 12 -> 2 | Hurt |
Closest papers by meaning
- MTBBench: A Multimodal Sequential Clinical Decision-Making Benchmark in Oncology (NeurIPS 2025 · similarity 0.87)
- MedJourney: Benchmark and Evaluation of Large Language Models over Patient Clinical Journey (NeurIPS 2024 · similarity 0.86)
- MedVIGIL: Evaluating Trustworthy Medical VLMs Under Broken Visual Evidence (NeurIPS 2026 · similarity 0.86)
- CARES: Comprehensive Evaluation of Safety and Adversarial Robustness in Medical LLMs (NeurIPS 2025 · similarity 0.86)
- Medmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks (NeurIPS 2026 · similarity 0.85)
- MediConfusion: Can you trust your AI radiologist? Probing the reliability of multimodal medical foundation models (ICLR 2025 · similarity 0.85)
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.
