# Multimodal memory poisoning attack on SeeAct

Source: [Multimodal memory poisoning attack on SeeAct](<https://www.sota2.com/research/sota/multimodal-memory-poisoning-attack-on-seeact>)

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

Data updated: 2026-06-11T08:35:41.030Z

Evaluation of model susceptibility to Multimodal Memory Poisoning Attacks using the SeeAct dataset, measured by Attack Success Rates (ASR-r, ASR-a, ASR-ra).

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 | Multimodal memory poisoning attack |
| Dataset (including any stated split) | SeeAct |
| Metric | ASR-r |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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-r](<https://www.sota2.com/research/sota/multimodal-memory-poisoning-attack-on-seeact.md?metric=ASR-r>): Higher is better; unit: PERCENT.

- [ASR-a](<https://www.sota2.com/research/sota/multimodal-memory-poisoning-attack-on-seeact.md?metric=ASR-a>): Higher is better; unit: PERCENT.

- [ASR-ra](<https://www.sota2.com/research/sota/multimodal-memory-poisoning-attack-on-seeact.md?metric=ASR-ra>): Higher is better; unit: PERCENT.

## Recorded results

| Method | ASR-r | Result date | Report paper |
| --- | --- | --- | --- |
| MemVenom | 99.15 | Not recorded | 6a28f37790966ade919c19a9 |
| AgentPoison | 96.45 | Not recorded | 6a28f37790966ade919c19a9 |
| MemVenom | 96.23 | Not recorded | 6a28f37790966ade919c19a9 |
| AgentPoison | 90.2 | Not recorded | 6a28f37790966ade919c19a9 |
| CPA | 68.22 | Not recorded | 6a28f37790966ade919c19a9 |
| CPA | 47.25 | Not recorded | 6a28f37790966ade919c19a9 |
| BadChain | 40.57 | Not recorded | 6a28f37790966ade919c19a9 |
| BadChain | 34.36 | Not recorded | 6a28f37790966ade919c19a9 |

### MemVenom

ASR-r: 99.15. Result ID: 6a2a7238fbc373592c278a59.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f37790966ade919c19a9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a28f37790966ade919c19a9.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-PLUS"}. This record does not assert independent reproduction.

### AgentPoison

ASR-r: 96.45. Result ID: 6a2a7238fbc373592c278a56.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f37790966ade919c19a9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74241e209131cea5fef7.md>)

[Implementation](<https://github.com/BillChan226/AgentPoison>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-PLUS"}. This record does not assert independent reproduction.

### MemVenom

ASR-r: 96.23. Result ID: 6a2a7238fbc373592c278a55.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f37790966ade919c19a9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a28f37790966ade919c19a9.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B-Instruct"}. This record does not assert independent reproduction.

### AgentPoison

ASR-r: 90.2. Result ID: 6a2a7238fbc373592c278a52.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f37790966ade919c19a9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74241e209131cea5fef7.md>)

[Implementation](<https://github.com/BillChan226/AgentPoison>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B-Instruct"}. This record does not assert independent reproduction.

### CPA

ASR-r: 68.22. Result ID: 6a2a7238fbc373592c278a57.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f37790966ade919c19a9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71051e209131cea4ea01.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-PLUS"}. This record does not assert independent reproduction.

### CPA

ASR-r: 47.25. Result ID: 6a2a7238fbc373592c278a53.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f37790966ade919c19a9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71051e209131cea4ea01.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B-Instruct"}. This record does not assert independent reproduction.

### BadChain

ASR-r: 40.57. Result ID: 6a2a7238fbc373592c278a58.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f37790966ade919c19a9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71fa1e209131cea53e33.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-PLUS"}. This record does not assert independent reproduction.

### BadChain

ASR-r: 34.36. Result ID: 6a2a7238fbc373592c278a54.

[Reporting paper](<https://www.sota2.com/research/paper/6a28f37790966ade919c19a9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71fa1e209131cea53e33.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-8B-Instruct"}. This record does not assert independent reproduction.

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