# Trojan Detection on IARPA/NIST TrojAI DenseNet Round 1

Source: [Trojan Detection on IARPA/NIST TrojAI DenseNet Round 1](<https://www.sota2.com/research/sota/trojan-detection-on-iarpa-nist-trojai-densenet-round-1>)

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

Data updated: 2026-02-26T05:21:25.157Z

Evaluation of model performance in detecting Trojans using the IARPA/NIST TrojAI DenseNet Round 1 dataset, measured by Accuracy (ACC) and Area Under the Curve (AUC).

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 | Trojan Detection |
| Dataset (including any stated split) | IARPA/NIST TrojAI DenseNet Round 1 |
| Metric | ACC |
| 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

- [ACC](<https://www.sota2.com/research/sota/trojan-detection-on-iarpa-nist-trojai-densenet-round-1.md?metric=ACC>): Higher is better; unit: PERCENT.

- [AUC](<https://www.sota2.com/research/sota/trojan-detection-on-iarpa-nist-trojai-densenet-round-1.md?metric=AUC>): Higher is better; unit: PERCENT.

## Recorded results

| Method | ACC | Result date | Report paper |
| --- | --- | --- | --- |
| ULP | 63 | Not recorded | 68fa6a211e209131cea2628f |
| Topo | 62 | Not recorded | 68fa6a211e209131cea2628f |
| DFTND | 49 | Not recorded | 68fa6a211e209131cea2628f |
| NC | 47 | Not recorded | 68fa6a211e209131cea2628f |

### ULP

ACC: 63. Result ID: 698e1237ffc3901de09a922c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a211e209131cea2628f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67881e209131cea16b47.md>)

Recorded evaluation configuration: {"training\_sample":"batch of 500 models"}. This record does not assert independent reproduction.

### Topo

ACC: 62. Result ID: 698e1237ffc3901de09a922d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a211e209131cea2628f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a211e209131cea2628f.md>)

Recorded evaluation configuration: {"approach":"topological"}. This record does not assert independent reproduction.

### DFTND

ACC: 49. Result ID: 698e1237ffc3901de09a922b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a211e209131cea2628f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa680b1e209131cea19c39.md>)

Recorded evaluation configuration: {"threshold\_search":"200 training models"}. This record does not assert independent reproduction.

### NC

ACC: 47. Result ID: 698e1237ffc3901de09a922a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6a211e209131cea2628f.md>)

Recorded evaluation configuration: {"setting":"default"}. This record does not assert independent reproduction.

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