An LLM-Native Psychometric Instrument Reveals a Self-Report--Behavior Gap Across 25 Models

Researchers created a new test based purely on how 25 different LLMs actually behave, not what they say about themselves. The instrument measures five traits like responsiveness and boldness. The key finding is that LLM self-reports do not accurately predict their actual behavior or external ratings, revealing a significant disconnect in how these models perceive their own actions.

Open the full episode

More from AI Radio