Risk Evaluation · Study

"Did you lie?": Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms

NUWA research note · independent review pending

First published: 2026-06-17Source checked: 2026-10-25

Original source ↗Cite this record ↓Deception →

NUWA Research Note

Research question Design and assumptions Source findings NUWA interpretation Limits and contrary evidence Research and decision relevance Institutions and roles Original resources

Model-assisted reading · Selected sections of the original full text or official article; a complete section-by-section audit remains pending.

At a glance

NUWA interpretation: testbed belief validity and selection can explain apparent detector strength.

Research question

Do lie detectors generalize when testbeds verify that stated answers contradict model beliefs?

Design and assumptions

Thirteen trained organisms and prompted lying across 31 open models evaluate four detector families; belief verification uses chain-of-thought and held-out behavior.

Source findings

Activation/logprob detectors degraded on trained organisms.

The CoT judge retained 0.82 balanced accuracy, partly because belief verification selects CoT-readable cases.

NUWA interpretation

NUWA interpretation: testbed belief validity and selection can explain apparent detector strength.

Limits and contrary evidence

Belief verification is probabilistic, and the CoT-readable selection favors the CoT judge; results do not certify internal beliefs.

Research and decision relevance

Model recommendation: report belief-validation and selection assumptions alongside detection accuracy.

Precise source locations

“Did you lie?” Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms · AbstractSelected original resource https://arxiv.org/html/2606.12618; sha256:22f4125cdf9db0f78a5cd6677f022405de2a3021c8a46fa892dc485e2e15c240; archived source capture; exact capture time not independently established
“Did you lie?” Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms · §4.2 Model OrganismsSelected original resource https://arxiv.org/html/2606.12618; sha256:22f4125cdf9db0f78a5cd6677f022405de2a3021c8a46fa892dc485e2e15c240; archived source capture; exact capture time not independently established
“Did you lie?” Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms · §6 ResultsSelected original resource https://arxiv.org/html/2606.12618; sha256:22f4125cdf9db0f78a5cd6677f022405de2a3021c8a46fa892dc485e2e15c240; archived source capture; exact capture time not independently established
“Did you lie?” Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms · §8 LimitationsSelected original resource https://arxiv.org/html/2606.12618; sha256:22f4125cdf9db0f78a5cd6677f022405de2a3021c8a46fa892dc485e2e15c240; archived source capture; exact capture time not independently established

Source checking, extraction, editorial approval and replication are recorded separately. Methodology →

Institutions and roles

Published by UK AI Security Institute

Authors (3)

Alan Cooney · David Africa · Geoffrey Irving

Cite and reuse

Cite the original publication for its findings. Cite this NUWA record separately for its explanation and retained snapshot.

BibTeX ↓RIS ↓Record + source metadata · JSON ↓Reading coverage · JSON Suggest a correction

Original publication · BibTeX

@misc{belief-verified-lie-detectors,
  title = {"Did you lie?": Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms},
  url = {https://www.aisi.gov.uk/research/did-you-lie-evaluating-lie-detectors-across-model-scale-and-belief-verified-model-organisms},
  author = {Alan Cooney and David Africa and Geoffrey Irving},
  year = {2026},
  date = {2026-06-17},
  version = {Source capture 2026-10-25},
  note = {Source capture 2026-10-25}
}

NUWA record · BibTeX

@misc{belief-verified-lie-detectors-nuwa,
  title = {"Did you lie?": Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms — NUWA research record},
  url = {https://whitzard.tech/en/nuwa/whitzard-index/work/belief-verified-lie-detectors/},
  version = {2026-10-29.1},
  organization = {NUWA Frontier AI Safety Lab},
  howpublished = {Retained note edition: https://whitzard.tech/nuwa/whitzard-index/data/record-editions/2026-10-29.1--belief-verified-lie-detectors.json},
  note = {NUWA research note. Source-reading scope and independent review status are stated on the record. Snapshot 2026-10-29.1. Retained note edition: https://whitzard.tech/nuwa/whitzard-index/data/record-editions/2026-10-29.1--belief-verified-lie-detectors.json.}
}

Evidence snapshot 2026-10-29.1 · Index updates · RSS

References

[1] Alan Cooney, David Africa, et al.. "Did you lie?": Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms. 2026-06-17 ; Official publication masthead and article byline
[2] Alan Cooney, David Africa, et al.. “Did you lie?” Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organisms ; Abstract; §4.2 Model Organisms; §6 Results