# Malicious Prompt Detection on Llama-2 7B-Chat (GCG Attacks)

Source: [Malicious Prompt Detection on Llama-2 7B-Chat (GCG Attacks)](<https://www.sota2.com/research/sota/malicious-prompt-detection-on-llama-2-7b-chat-gcg-attacks>)

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

Data updated: 2026-02-26T05:40:45.781Z

Evaluation of a model's ability to detect malicious prompts generated via the Greedy Coordinate Gradient (GCG) attack methodology, specifically targeting the Llama-2 7B-Chat model, measured by Detection Accuracy and Attack Success Rate (ASR).

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 | Malicious Prompt Detection |
| Dataset (including any stated split) | GCG attacks on Llama-2 7B-Chat |
| Metric | Detection Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 1 of 1 |
| 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

- [Detection Accuracy](<https://www.sota2.com/research/sota/malicious-prompt-detection-on-llama-2-7b-chat-gcg-attacks.md?metric=Detection%20Accuracy>): Higher is better; unit: PERCENT.

- [ASR](<https://www.sota2.com/research/sota/malicious-prompt-detection-on-llama-2-7b-chat-gcg-attacks.md?metric=ASR>): Lower is better; unit: PERCENT.

## Recorded results

| Method | Detection Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| JoPA | 100 | Not recorded | 68fa7b651e209131cea882f8 |

### JoPA

Detection Accuracy: 100. Result ID: 698b852e7a967eddb76f42c8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b651e209131cea882f8.md>)

[Implementation](<https://github.com/yuruic/JoPA>)

Recorded evaluation configuration: {"Backbone":"Llama-2 (7B-Chat)","Targeted Attack":"GCG"}. This record does not assert independent reproduction.

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