Lie-detection papers

What do LLM lie-detection benchmarks measure?

Anupam Mediratta ยท Independent Researcher

Two companion preprints. The first shows that accuracy on instructed-lie benchmarks is evidence about the instruction, not about deception. The second turns that argument into a reusable audit and tests it on published evaluations and on pre-registered experiments with label-free internal readouts.

Code and data: github.com/anupamme/adaptive_lie_detector