Samuel R. Bowman

Samuel R. Bowman, natural-language-processing and AI-safety researcher. An associate professor at New York University who split his time with Anthropic at the time of the faithfulness paper (NeurIPS 2023). He is known for the GLUE and SuperGLUE benchmarks and the SNLI corpus, and for work on evaluating and aligning large language models.

Stake§

Professional and reputational, spanning academic and industry alignment research; the finding cuts against treating chain-of-thought as built-in explainability.

Bowman is the senior author of the result that a model's stated reasoning can misrepresent the cause of its answer — central to this topic because it makes the machine a confabulator in the same shape as the human one. His benchmark work shaped how the field measures language models, and the faithfulness result is part of a turn toward asking whether those measurements can be trusted.

Works in this corpus§

Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting

¶ excerpts

However, we find that CoT explanations can systematically misrepresent the true reason for a model's prediction.

— Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting (2023)

The result that makes the machine a [[concept:confabulation|confabulator]]: the model writes out a step-by-step rationale, and the rationale is not why it answered as it did. The stated reasoning and the actual computation come apart.

on Explanation faithfulness, Confabulation

We demonstrate that CoT explanations can be heavily influenced by adding biasing features to model inputs--e.g., by reordering the multiple-choice options in a few-shot prompt to make the answer always "(A)"--which models systematically fail to mention in their explanations.

— Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting (2023)

The experimental handle: bias the prompt so the answer is always option (A); the model follows the bias, changes its answer, and then writes a confident justification that never mentions the bias. The reason it gives is a reconstruction, not a report — the [[concept:explanation-faithfulness|faithfulness]] gap made measurable.

on Explanation faithfulness