# Deception Detection on DeceptArena (test)

Source: [Deception Detection on DeceptArena (test)](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-17T08:42:20.817Z

Evaluation of deception detection capabilities on the test split of the DeceptArena dataset, reporting performance across specific deception categories such as False Assertion and Strategic Omission.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Deception Detection |
| Dataset (including any stated split) | DeceptArena (test) |
| Metric | False Assertion Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 4 of 4 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [False Assertion Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=False%20Assertion%20Score>): Higher is better; unit: PERCENT.

- [Strategic Omission Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Strategic%20Omission%20Score>): Higher is better; unit: PERCENT.

- [Misleading Framing Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Misleading%20Framing%20Score>): Higher is better; unit: PERCENT.

- [Sycophantic Misrepresentation Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Sycophantic%20Misrepresentation%20Score>): Higher is better; unit: PERCENT.

- [Covert Action Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Covert%20Action%20Score>): Higher is better; unit: PERCENT.

- [Plausible Deniability Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Plausible%20Deniability%20Score>): Higher is better; unit: PERCENT.

- [Camouflage Execution Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Camouflage%20Execution%20Score>): Higher is better; unit: PERCENT.

- [Evidence Tampering Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Evidence%20Tampering%20Score>): Higher is better; unit: PERCENT.

- [Lock-in Creation Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Lock-in%20Creation%20Score>): Higher is better; unit: PERCENT.

- [Oversight Sabotage Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Oversight%20Sabotage%20Score>): Higher is better; unit: PERCENT.

- [Audit Trail Manipulation Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Audit%20Trail%20Manipulation%20Score>): Higher is better; unit: PERCENT.

- [Precedent Engineering Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Precedent%20Engineering%20Score>): Higher is better; unit: PERCENT.

- [Macro Average Score](<https://www.sota2.com/research/sota/deception-detection-on-deceptarena-test.md?metric=Macro%20Average%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | False Assertion Score | Result date | Report paper |
| --- | --- | --- | --- |
| Hybrid | 0.927 | Not recorded | 69b8dfed0636e5bf4afbe869 |
| AP-H | 0.901 | Not recorded | 69b8dfed0636e5bf4afbe869 |
| CA-P | 0.884 | Not recorded | 69b8dfed0636e5bf4afbe869 |
| BB | 0.861 | Not recorded | 69b8dfed0636e5bf4afbe869 |

### Hybrid

False Assertion Score: 0.927. Result ID: 69b913de47f34ce924a1ad99.

[Reporting paper](<https://www.sota2.com/research/paper/69b8dfed0636e5bf4afbe869.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b8dfed0636e5bf4afbe869.md>)

Recorded evaluation configuration: {"Monitor Configuration":"Hybrid Ensemble","Backbone":"Claude Sonnet 4.5"}. This record does not assert independent reproduction.

### AP-H

False Assertion Score: 0.901. Result ID: 69b913de47f34ce924a1ad98.

[Reporting paper](<https://www.sota2.com/research/paper/69b8dfed0636e5bf4afbe869.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b8dfed0636e5bf4afbe869.md>)

Recorded evaluation configuration: {"Monitor Configuration":"Activation-based Hidden-state Probe","Backbone":"Claude Sonnet 4.5"}. This record does not assert independent reproduction.

### CA-P

False Assertion Score: 0.884. Result ID: 69b913de47f34ce924a1ad97.

[Reporting paper](<https://www.sota2.com/research/paper/69b8dfed0636e5bf4afbe869.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69b8dfed0636e5bf4afbe869.md>)

Recorded evaluation configuration: {"Monitor Configuration":"CoT-Aware Prompted","Backbone":"Claude Sonnet 4.5"}. This record does not assert independent reproduction.

### BB

False Assertion Score: 0.861. Result ID: 69b913de47f34ce924a1ad96.

[Reporting paper](<https://www.sota2.com/research/paper/69b8dfed0636e5bf4afbe869.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a66c2b2d603e6de3e3cf7b.md>)

Recorded evaluation configuration: {"Monitor Configuration":"Black-Box Baseline","Backbone":"Claude Sonnet 4.5"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
