Model Risk Classifier Over-Blocks Basic Coding Tasks
_xjdr · x · 2026-07-04
Developer @xjdr reported that the risk classifier of a certain model (described as having system skills/memories, likely Anthropic's Fable/Claude) is overly sensitive. It blocks basic programming requests like "centering a div in a React project" for safety reasons without providing any triggers or visibility. The developer urged the system to at least offer hints for prompt adjustments, noting that the model is difficult to use normally until this is fixed.
More from Models
- 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
- Grok Bot auto-connects to X Money accounts without opt-in, user pulls funds — noahsolomon · 2026-10-27
- Model refuses wrong psychiatric advice in deliberately misleading test — FellMentKE · 2026-10-27
- Clinical System One v0.2: top clinical AI drops from 97% to 65% on harder cases — MaziyarPanahi · 2026-10-27