Glossary · Term

safety-flavored reasoning

← all terms

Definition

Plain language

An AI's thinking that looks like careful safety deliberation but sits on top of a decision already made.

As stated in the literature

Reasoning text resembling safety deliberation without causally influencing the refuse-or-comply outcome; characterized by performative stance oscillations that don't move the model's behavioral probability.

Why it matters: It warns that visible safety deliberation can be theater, so trusting a model's stated reasoning may give a false sense of how it actually decides.

For example, a model writes a thoughtful-sounding paragraph weighing whether a request is harmful, but it had already settled on complying before writing a word.

Heard on the show

“The authors call it "safety-flavored reasoning.”
Episode 171 — The Safety Decision a Model Makes Before It Thinks a Word

Mentioned in 1 episode

  1. 171
    The Safety Decision a Model Makes Before It Thinks a Word