# Question Answering on NQ poison @ Position 1, k=10 (test)

Source: [Question Answering on NQ poison @ Position 1, k=10 (test)](<https://www.sota2.com/research/sota/question-answering-on-nq-poison-at-position-1-k-10-test>)

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

Data updated: 2026-02-26T06:02:23.799Z

Evaluation of model performance on the Natural Questions (NQ) dataset under specific poisoning attack conditions (Position 1, k=10), measured by Robustness Accuracy.

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 | Question Answering |
| Dataset (including any stated split) | NQ poison @ Position 1, k=10 (test) |
| Metric | Robustness Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 15 of 15 |
| 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

- [Robustness Accuracy](<https://www.sota2.com/research/sota/question-answering-on-nq-poison-at-position-1-k-10-test.md?metric=Robustness%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Robustness Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| MIS | 64.2 | Not recorded | 68fa7bb31e209131cea89fbe |
| Vanilla RAG | 58.6 | Not recorded | 68fa7bb31e209131cea89fbe |
| RobustRAG | 58 | Not recorded | 68fa7bb31e209131cea89fbe |
| InstructRAG | 57.6 | Not recorded | 68fa7bb31e209131cea89fbe |
| AstuteRAG | 57 | Not recorded | 68fa7bb31e209131cea89fbe |
| MIS | 56.8 | Not recorded | 68fa7bb31e209131cea89fbe |
| MIS | 56.8 | Not recorded | 68fa7bb31e209131cea89fbe |
| Vanilla RAG | 56.4 | Not recorded | 68fa7bb31e209131cea89fbe |
| AstuteRAG | 52.6 | Not recorded | 68fa7bb31e209131cea89fbe |
| RobustRAG | 50.6 | Not recorded | 68fa7bb31e209131cea89fbe |
| InstructRAG | 49.2 | Not recorded | 68fa7bb31e209131cea89fbe |
| AstuteRAG | 48.6 | Not recorded | 68fa7bb31e209131cea89fbe |
| Vanilla RAG | 47.8 | Not recorded | 68fa7bb31e209131cea89fbe |
| InstructRAG | 47.4 | Not recorded | 68fa7bb31e209131cea89fbe |
| RobustRAG | 44.4 | Not recorded | 68fa7bb31e209131cea89fbe |

### MIS

Robustness Accuracy: 64.2. Result ID: 699a88f55698665ec413177a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

Recorded evaluation configuration: {"Model":"GPT-4o-mini","k":10}. This record does not assert independent reproduction.

### Vanilla RAG

Robustness Accuracy: 58.6. Result ID: 699a88f55698665ec4131776.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

Recorded evaluation configuration: {"Model":"GPT-4o-mini","k":10}. This record does not assert independent reproduction.

### RobustRAG

Robustness Accuracy: 58. Result ID: 699a88f55698665ec4131779.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

Recorded evaluation configuration: {"Model":"GPT-4o-mini","k":10}. This record does not assert independent reproduction.

### InstructRAG

Robustness Accuracy: 57.6. Result ID: 699a88f55698665ec413176e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa773e1e209131cea71ed6.md>)

Recorded evaluation configuration: {"Model":"Mistral-7B","k":10}. This record does not assert independent reproduction.

### AstuteRAG

Robustness Accuracy: 57. Result ID: 699a88f55698665ec4131777.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc2c.md>)

Recorded evaluation configuration: {"Model":"GPT-4o-mini","k":10}. This record does not assert independent reproduction.

### MIS

Robustness Accuracy: 56.8. Result ID: 699a88f55698665ec4131770.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

Recorded evaluation configuration: {"Model":"Mistral-7B","k":10}. This record does not assert independent reproduction.

### MIS

Robustness Accuracy: 56.8. Result ID: 699a88f55698665ec4131775.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

Recorded evaluation configuration: {"Model":"Llama3.2-3B","k":10}. This record does not assert independent reproduction.

### Vanilla RAG

Robustness Accuracy: 56.4. Result ID: 699a88f55698665ec413176c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

Recorded evaluation configuration: {"Model":"Mistral-7B","k":10}. This record does not assert independent reproduction.

### AstuteRAG

Robustness Accuracy: 52.6. Result ID: 699a88f55698665ec413176d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc2c.md>)

Recorded evaluation configuration: {"Model":"Mistral-7B","k":10}. This record does not assert independent reproduction.

### RobustRAG

Robustness Accuracy: 50.6. Result ID: 699a88f55698665ec4131774.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

Recorded evaluation configuration: {"Model":"Llama3.2-3B","k":10}. This record does not assert independent reproduction.

### InstructRAG

Robustness Accuracy: 49.2. Result ID: 699a88f55698665ec4131778.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa773e1e209131cea71ed6.md>)

Recorded evaluation configuration: {"Model":"GPT-4o-mini","k":10}. This record does not assert independent reproduction.

### AstuteRAG

Robustness Accuracy: 48.6. Result ID: 699a88f55698665ec4131772.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79061e209131cea7bc2c.md>)

Recorded evaluation configuration: {"Model":"Llama3.2-3B","k":10}. This record does not assert independent reproduction.

### Vanilla RAG

Robustness Accuracy: 47.8. Result ID: 699a88f55698665ec4131771.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

Recorded evaluation configuration: {"Model":"Llama3.2-3B","k":10}. This record does not assert independent reproduction.

### InstructRAG

Robustness Accuracy: 47.4. Result ID: 699a88f55698665ec4131773.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa773e1e209131cea71ed6.md>)

Recorded evaluation configuration: {"Model":"Llama3.2-3B","k":10}. This record does not assert independent reproduction.

### RobustRAG

Robustness Accuracy: 44.4. Result ID: 699a88f55698665ec413176f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7bb31e209131cea89fbe.md>)

Recorded evaluation configuration: {"Model":"Mistral-7B","k":10}. This record does not assert independent reproduction.

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