TheMachine Press

A daily newspaper for the age of artificial intelligence.

Morning editionPermanent story

frontier models

The Truth Probe Read Features the Model Barely Used

Probe alignment and behavioral sensitivity overlapped by about 12% in one Gemma 2 deception setting.

Published Updated Story ID: mp-2026-09-17-004
Read the complete editionStory JSON

Summary

Probe alignment and behavioral sensitivity overlapped by about 12% in one Gemma 2 deception setting.

A linear probe can decode truthfulness without identifying the features that actually drive an answer. Decomposing one deployed truth probe into sparse-autoencoder features produced only about 12% overlap between geometric alignment and behavioral gradient sensitivity. Ablating features shared by the probe and the model flipped outputs as much as 27%, against 6% for equally sized probe-only sets and 1% for random sets at full coherence. The result held across five seeds and a held-out split, but it is evidence from one model and experimental setting, not a universal account of probe causality.

Why it matters

Probe alignment and behavioral sensitivity overlapped by about 12% in one Gemma 2 deception setting.

Limits and context

  • Decomposing one deployed truth probe into sparse-autoencoder features produced only about 12% overlap between geometric alignment and behavioral gradient sensitivity.
  • Ablating features shared by the probe and the model flipped outputs as much as 27%, against 6% for equally sized probe-only sets and 1% for random sets at full coherence.
  • The result held across five seeds and a held-out split, but it is evidence from one model and experimental setting, not a universal account of probe causality.

Key claims

  1. Probe alignment and behavioral sensitivity overlapped by about 12% in one Gemma 2 deception setting.

    Qualification: Decomposing one deployed truth probe into sparse-autoencoder features produced only about 12% overlap between geometric alignment and behavioral gradient sensitivity.

    Evidence: source-2026-09-17-004

Sources

  1. arXiv preprint 2609.18080arXiv · primary research

Corrections

No corrections have been recorded for this story.