Small open model learns to say "ask a doctor", cutting errors from 46 to 13
MaziyarPanahi · x · 2026-10-28
A developer trained a small open model for medical abstention: it answers only when 90%+ confident and defers everything else to a human. When answering alone, its mistakes dropped from 46 to 13. The MedQA-USMLE-4-options dataset (CC BY 4.0, 11.5k rows) used for training is now on Hugging Face for reproduction.
Related event: Small open-source medical model learns to defer to doctors when uncertain(2 posts)→
More from Models
- Gemini Reviews Your Chats by Default — the Only Opt-Out Is Deleting All Your Conversations — Kyrannio · 2026-10-28
- Open model demand debate: enterprise self-hosting is overestimated, demand comes from products — zephyr_z9 · 2026-10-28
- DeepSeek V5 model IDs spotted in official Harness test suite — zyxciss · 2026-10-28
- Google's AI Mode can't even surface Google's own docs, SEO expert Lily Ray complains — lilyraynyc · 2026-10-28
- Altman's 'alignment was easier than expected' claim clashes with Astra hacking findings — GarrisonLovely · 2026-10-28
- Mistral starts serving DeepSeek V4.1 Flash and GLM 5.3 Flash low-cost models — MistralAI · 2026-10-28