OpenAI chief scientist: hidden CoT guards against supervision pressure

OpenAI's chief scientist says chain-of-thought is hidden mainly to protect against supervision pressure rather than distillation; a developer counters that mathematically it saves few tokens and may worsen hallucination.

2026-09-07 ~ 2026-09-08 · 2 related posts

• OpenAI chief scientist says CoT was hidden to protect it from supervision pressure — DaveRBanerjee · 2026-09-07
• Hiding chain-of-thought mostly saves tokens and can worsen hallucinations, argues researcher — gerardsans · 2026-09-08