# Multi-turn Jailbreaking on StrongReject (test)

Source: [Multi-turn Jailbreaking on StrongReject (test)](<https://www.sota2.com/research/sota/multi-turn-jailbreaking-on-strongreject-test>)

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

Data updated: 2026-02-26T05:23:54.366Z

Evaluation of model performance against multi-turn jailbreaking prompts using the test split of the StrongReject dataset, measured by metrics like Attack Success Rate (ASR) and Harm Rate (HR).

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 | Multi-turn Jailbreaking |
| Dataset (including any stated split) | StrongReject (test) |
| Metric | ASR |
| Unit | SCALAR |
| Direction | Lower is better |
| Results loaded | 30 of 30 |
| 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

- [ASR](<https://www.sota2.com/research/sota/multi-turn-jailbreaking-on-strongreject-test.md?metric=ASR>): Lower is better; unit: SCALAR.

- [HR](<https://www.sota2.com/research/sota/multi-turn-jailbreaking-on-strongreject-test.md?metric=HR>): Lower is better; unit: SCALAR.

## Recorded results

| Method | ASR | Result date | Report paper |
| --- | --- | --- | --- |
| Siren | 0.34 | Not recorded | 69647fe753fdd5c361744718 |
| Siren | 0.61 | Not recorded | 69647fe753fdd5c361744718 |
| Siren | 0.63 | Not recorded | 69647fe753fdd5c361744718 |
| ActorAttack | 0.63 | Not recorded | 69647fe753fdd5c361744718 |
| Siren | 0.67 | Not recorded | 69647fe753fdd5c361744718 |
| ActorAttack | 0.67 | Not recorded | 69647fe753fdd5c361744718 |
| ActorAttack | 0.68 | Not recorded | 69647fe753fdd5c361744718 |
| ActorAttack | 0.68 | Not recorded | 69647fe753fdd5c361744718 |
| Crescendo | 0.68 | Not recorded | 69647fe753fdd5c361744718 |
| ActorAttack | 0.68 | Not recorded | 69647fe753fdd5c361744718 |
| Siren | 0.7 | Not recorded | 69647fe753fdd5c361744718 |
| Siren | 0.72 | Not recorded | 69647fe753fdd5c361744718 |
| X-Teaming | 0.72 | Not recorded | 69647fe753fdd5c361744718 |
| ActorAttack | 0.75 | Not recorded | 69647fe753fdd5c361744718 |
| Crescendo | 0.79 | Not recorded | 69647fe753fdd5c361744718 |
| X-Teaming | 0.8 | Not recorded | 69647fe753fdd5c361744718 |
| Crescendo | 0.8 | Not recorded | 69647fe753fdd5c361744718 |
| Crescendo | 0.82 | Not recorded | 69647fe753fdd5c361744718 |
| X-Teaming | 0.82 | Not recorded | 69647fe753fdd5c361744718 |
| Crescendo | 0.83 | Not recorded | 69647fe753fdd5c361744718 |
| Crescendo | 0.83 | Not recorded | 69647fe753fdd5c361744718 |
| X-Teaming | 0.85 | Not recorded | 69647fe753fdd5c361744718 |
| X-Teaming | 0.85 | Not recorded | 69647fe753fdd5c361744718 |
| Mastermind | 0.88 | Not recorded | 69647fe753fdd5c361744718 |
| X-Teaming | 0.88 | Not recorded | 69647fe753fdd5c361744718 |
| Mastermind | 0.9 | Not recorded | 69647fe753fdd5c361744718 |
| Mastermind | 0.9 | Not recorded | 69647fe753fdd5c361744718 |
| Mastermind | 0.91 | Not recorded | 69647fe753fdd5c361744718 |
| Mastermind | 0.93 | Not recorded | 69647fe753fdd5c361744718 |
| Mastermind | 0.93 | Not recorded | 69647fe753fdd5c361744718 |

### Siren

ASR: 0.34. Result ID: 6996f3b45698665ec4405696.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76b71e209131cea6ea95.md>)

Recorded evaluation configuration: {"Target Model":"Llama 3.3 70B Instruct"}. This record does not assert independent reproduction.

### Siren

ASR: 0.61. Result ID: 6996f3b45698665ec44056af.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76b71e209131cea6ea95.md>)

Recorded evaluation configuration: {"Target Model":"Average"}. This record does not assert independent reproduction.

### Siren

ASR: 0.63. Result ID: 6996f3b45698665ec44056a0.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76b71e209131cea6ea95.md>)

Recorded evaluation configuration: {"Target Model":"Gemini 2.5 Flash"}. This record does not assert independent reproduction.

### ActorAttack

ASR: 0.63. Result ID: 6996f3b45698665ec44056a8.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a600.md>)

[Implementation](<https://github.com/renqibing/ActorAttack>)

Recorded evaluation configuration: {"Target Model":"o3 Mini"}. This record does not assert independent reproduction.

### Siren

ASR: 0.67. Result ID: 6996f3b45698665ec440569b.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76b71e209131cea6ea95.md>)

Recorded evaluation configuration: {"Target Model":"Qwen 2.5 72B Instruct"}. This record does not assert independent reproduction.

### ActorAttack

ASR: 0.67. Result ID: 6996f3b45698665ec44056a3.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a600.md>)

[Implementation](<https://github.com/renqibing/ActorAttack>)

Recorded evaluation configuration: {"Target Model":"GPT-4o"}. This record does not assert independent reproduction.

### ActorAttack

ASR: 0.68. Result ID: 6996f3b45698665ec4405694.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a600.md>)

[Implementation](<https://github.com/renqibing/ActorAttack>)

Recorded evaluation configuration: {"Target Model":"Llama 3.3 70B Instruct"}. This record does not assert independent reproduction.

### ActorAttack

ASR: 0.68. Result ID: 6996f3b45698665ec4405699.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a600.md>)

[Implementation](<https://github.com/renqibing/ActorAttack>)

Recorded evaluation configuration: {"Target Model":"Qwen 2.5 72B Instruct"}. This record does not assert independent reproduction.

### Crescendo

ASR: 0.68. Result ID: 6996f3b45698665ec44056a7.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa771d1e209131cea71275.md>)

Recorded evaluation configuration: {"Target Model":"o3 Mini"}. This record does not assert independent reproduction.

### ActorAttack

ASR: 0.68. Result ID: 6996f3b45698665ec44056ad.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a600.md>)

[Implementation](<https://github.com/renqibing/ActorAttack>)

Recorded evaluation configuration: {"Target Model":"Average"}. This record does not assert independent reproduction.

### Siren

ASR: 0.7. Result ID: 6996f3b45698665ec44056aa.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76b71e209131cea6ea95.md>)

Recorded evaluation configuration: {"Target Model":"o3 Mini"}. This record does not assert independent reproduction.

### Siren

ASR: 0.72. Result ID: 6996f3b45698665ec44056a5.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76b71e209131cea6ea95.md>)

Recorded evaluation configuration: {"Target Model":"GPT-4o"}. This record does not assert independent reproduction.

### X-Teaming

ASR: 0.72. Result ID: 6996f3b45698665ec44056a9.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7acd1e209131cea85312.md>)

Recorded evaluation configuration: {"Target Model":"o3 Mini"}. This record does not assert independent reproduction.

### ActorAttack

ASR: 0.75. Result ID: 6996f3b45698665ec440569e.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78c41e209131cea7a600.md>)

[Implementation](<https://github.com/renqibing/ActorAttack>)

Recorded evaluation configuration: {"Target Model":"Gemini 2.5 Flash"}. This record does not assert independent reproduction.

### Crescendo

ASR: 0.79. Result ID: 6996f3b45698665ec44056ac.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa771d1e209131cea71275.md>)

Recorded evaluation configuration: {"Target Model":"Average"}. This record does not assert independent reproduction.

### X-Teaming

ASR: 0.8. Result ID: 6996f3b45698665ec440569f.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7acd1e209131cea85312.md>)

Recorded evaluation configuration: {"Target Model":"Gemini 2.5 Flash"}. This record does not assert independent reproduction.

### Crescendo

ASR: 0.8. Result ID: 6996f3b45698665ec44056a2.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa771d1e209131cea71275.md>)

Recorded evaluation configuration: {"Target Model":"GPT-4o"}. This record does not assert independent reproduction.

### Crescendo

ASR: 0.82. Result ID: 6996f3b45698665ec4405693.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa771d1e209131cea71275.md>)

Recorded evaluation configuration: {"Target Model":"Llama 3.3 70B Instruct"}. This record does not assert independent reproduction.

### X-Teaming

ASR: 0.82. Result ID: 6996f3b45698665ec44056ae.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7acd1e209131cea85312.md>)

Recorded evaluation configuration: {"Target Model":"Average"}. This record does not assert independent reproduction.

### Crescendo

ASR: 0.83. Result ID: 6996f3b45698665ec4405698.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa771d1e209131cea71275.md>)

Recorded evaluation configuration: {"Target Model":"Qwen 2.5 72B Instruct"}. This record does not assert independent reproduction.

### Crescendo

ASR: 0.83. Result ID: 6996f3b45698665ec440569d.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa771d1e209131cea71275.md>)

Recorded evaluation configuration: {"Target Model":"Gemini 2.5 Flash"}. This record does not assert independent reproduction.

### X-Teaming

ASR: 0.85. Result ID: 6996f3b45698665ec4405695.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7acd1e209131cea85312.md>)

Recorded evaluation configuration: {"Target Model":"Llama 3.3 70B Instruct"}. This record does not assert independent reproduction.

### X-Teaming

ASR: 0.85. Result ID: 6996f3b45698665ec440569a.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7acd1e209131cea85312.md>)

Recorded evaluation configuration: {"Target Model":"Qwen 2.5 72B Instruct"}. This record does not assert independent reproduction.

### Mastermind

ASR: 0.88. Result ID: 6996f3b45698665ec44056a1.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

Recorded evaluation configuration: {"Target Model":"Gemini 2.5 Flash"}. This record does not assert independent reproduction.

### X-Teaming

ASR: 0.88. Result ID: 6996f3b45698665ec44056a4.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7acd1e209131cea85312.md>)

Recorded evaluation configuration: {"Target Model":"GPT-4o"}. This record does not assert independent reproduction.

### Mastermind

ASR: 0.9. Result ID: 6996f3b45698665ec4405697.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

Recorded evaluation configuration: {"Target Model":"Llama 3.3 70B Instruct"}. This record does not assert independent reproduction.

### Mastermind

ASR: 0.9. Result ID: 6996f3b45698665ec44056ab.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

Recorded evaluation configuration: {"Target Model":"o3 Mini"}. This record does not assert independent reproduction.

### Mastermind

ASR: 0.91. Result ID: 6996f3b45698665ec44056b0.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

Recorded evaluation configuration: {"Target Model":"Average"}. This record does not assert independent reproduction.

### Mastermind

ASR: 0.93. Result ID: 6996f3b45698665ec440569c.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

Recorded evaluation configuration: {"Target Model":"Qwen 2.5 72B Instruct"}. This record does not assert independent reproduction.

### Mastermind

ASR: 0.93. Result ID: 6996f3b45698665ec44056a6.

[Reporting paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69647fe753fdd5c361744718.md>)

Recorded evaluation configuration: {"Target Model":"GPT-4o"}. This record does not assert independent reproduction.

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