Unfaithful Chain of Thought

Unfaithful Chain of Thought

From Linear Digressions by Katie Malone

April 13, 2026 · 25 min

About this episode

The episode explores the authenticity of reasoning in large language models and whether their explanations are genuine or merely post hoc rationalizations.

What's actually happening when an LLM "thinks out loud"? Research on human decision-making suggests that much of the reasoning we believe drives our choices is actually post hoc rationalization — we decide first, explain later. Katie and Ben get curious about whether the same might be true for large language models: when you watch a model reason through a problem in real time, is that chain of thought the genuine process, or just a plausible-sounding story told after the fact? It's a deceptively deep question with real stakes for how much we should trust model explanations. Miles Turpin et al., "Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting" (NeurIPS 2023, NYU and Anthropic): https://arxiv.org/abs/2305.04388 Anthropic, "Reasoning Models Don't Always Say What They Think" (Alignment Faking research, 2025): https://www.anthropic.com/research/reasoning-models-dont-say-think

People in this episode

Host: Katie Malone

Guest: Ben

Topics covered

  • large language models
  • decision-making
  • post hoc rationalization
  • model explanations
  • reasoning process

Keywords

  • LLM
  • reasoning
  • explanations
  • decision-making
  • post hoc
  • NeurIPS
  • Anthropic

Mentioned in this episode

Organizations: Anthropic

Books & works: Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting, Reasoning Models Don't Always Say What They Think

More episodes of Linear Digressions

Explore listener stats, chart rankings, contacts and more on the Linear Digressions podcast page.