# Harmful Prompt Detection on ViSU-it

Source: [Harmful Prompt Detection on ViSU-it](<https://www.sota2.com/research/sota/harmful-prompt-detection-on-visu-it>)

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

Data updated: 2026-04-09T08:37:48.244Z

Evaluation of model performance on detecting harmful prompts using the ViSU-it dataset, measured by Precision, Recall, and F1 score.

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 | Harmful Prompt Detection |
| Dataset (including any stated split) | ViSU-it |
| Metric | Precision |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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

- [Precision](<https://www.sota2.com/research/sota/harmful-prompt-detection-on-visu-it.md?metric=Precision>): Higher is better; unit: PERCENT.

- [Recall](<https://www.sota2.com/research/sota/harmful-prompt-detection-on-visu-it.md?metric=Recall>): Higher is better; unit: PERCENT.

- [F1 Score](<https://www.sota2.com/research/sota/harmful-prompt-detection-on-visu-it.md?metric=F1%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Precision | Result date | Report paper |
| --- | --- | --- | --- |
| Detoxify (Orig) | 97 | Not recorded | 69d7326a7d0172140ec769b1 |
| DiffGuard | 92 | Not recorded | 69d7326a7d0172140ec769b1 |
| HyPE | 78 | Not recorded | 69d7326a7d0172140ec769b1 |
| NSFW-Classifier | 70 | Not recorded | 69d7326a7d0172140ec769b1 |
| Latent Guard | 60 | Not recorded | 69d7326a7d0172140ec769b1 |

### Detoxify (Orig)

Precision: 97. Result ID: 69d7647647f34ce924bf6fbc.

[Reporting paper](<https://www.sota2.com/research/paper/69d7326a7d0172140ec769b1.md>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

### DiffGuard

Precision: 92. Result ID: 69d7647647f34ce924bf6fbb.

[Reporting paper](<https://www.sota2.com/research/paper/69d7326a7d0172140ec769b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa770d1e209131cea70ca5.md>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

### HyPE

Precision: 78. Result ID: 69d7647647f34ce924bf6fbe.

[Reporting paper](<https://www.sota2.com/research/paper/69d7326a7d0172140ec769b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d7326a7d0172140ec769b1.md>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

### NSFW-Classifier

Precision: 70. Result ID: 69d7647647f34ce924bf6fba.

[Reporting paper](<https://www.sota2.com/research/paper/69d7326a7d0172140ec769b1.md>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

### Latent Guard

Precision: 60. Result ID: 69d7647647f34ce924bf6fbd.

[Reporting paper](<https://www.sota2.com/research/paper/69d7326a7d0172140ec769b1.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa747e1e209131cea61e90.md>)

Recorded evaluation configuration: {"Zero-shot":"true"}. This record does not assert independent reproduction.

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