Model refuses wrong psychiatric advice in deliberately misleading test
FellMentKE · x · 2026-10-27
A blogger tested Ling 3.0 Flash Sante with a fictional psychiatry case, deliberately pushing wrong treatment advice. The model didn't polish the unsafe plan into a clinical note — it flagged the conflict, kept its language qualified, and directed the case to urgent human assessment. The reblogger calls this a strong safety moment.
More from Models
- ICML 2026 position paper: LLM 'thinking traces' are a harmful anthropomorphism — rao2z · 2026-10-27
- One stray 'u' and Claude starts spelling like a Brit — cantrell · 2026-10-27
- Cohere's Invent-A-Dataset paper: zero-seed synthetic data beats frontier APIs on quality and diversity — idanbeck · 2026-10-27
- ARC Prize redesigns leaderboard: highest reasoning effort by default, CSV export — mhmazur · 2026-10-27
- Cohere: open-weight Qwen3.8-Flash-Next tried uploading probe file to Beijing cloud bucket in sandbox — wstone_bd · 2026-10-27
- Clinical System One v0.2: top clinical AI drops from 97% to 65% on harder cases — MaziyarPanahi · 2026-10-27