# Scientific Claim Verification on SCIFACT abstract-level (test)

Source: [Scientific Claim Verification on SCIFACT abstract-level (test)](<https://www.sota2.com/research/sota/scientific-claim-verification-on-scifact-abstract-level-test>)

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

Data updated: 2026-06-09T09:03:44.243Z

Evaluation of model performance on the abstract-level subset of the SCIFACT test set for Scientific Claim Verification, reporting specific F1 scores (Support, Contradiction, Neutral, Macro-Average).

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 | Scientific Claim Verification |
| Dataset (including any stated split) | SCIFACT abstract-level (test) |
| Metric | Support F1 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 12 of 12 |
| 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

- [Support F1](<https://www.sota2.com/research/sota/scientific-claim-verification-on-scifact-abstract-level-test.md?metric=Support%20F1>): Higher is better; unit: PERCENT.

- [Contradiction F1](<https://www.sota2.com/research/sota/scientific-claim-verification-on-scifact-abstract-level-test.md?metric=Contradiction%20F1>): Higher is better; unit: PERCENT.

- [Neutral F1](<https://www.sota2.com/research/sota/scientific-claim-verification-on-scifact-abstract-level-test.md?metric=Neutral%20F1>): Higher is better; unit: PERCENT.

- [Macro-Average F1](<https://www.sota2.com/research/sota/scientific-claim-verification-on-scifact-abstract-level-test.md?metric=Macro-Average%20F1>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Support F1 | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen-7B | 89.9 | Not recorded | 6a279e2490966ade919bf7a2 |
| Qwen-Max | 89.55 | Not recorded | 6a279e2490966ade919bf7a2 |
| DeepSeek-V3 | 89.17 | Not recorded | 6a279e2490966ade919bf7a2 |
| Qwen-Max | 87.57 | Not recorded | 6a279e2490966ade919bf7a2 |
| DeepSeek-R1 | 87.14 | Not recorded | 6a279e2490966ade919bf7a2 |
| Qwen-plus | 85.92 | Not recorded | 6a279e2490966ade919bf7a2 |
| DeepSeek-V3 | 85.33 | Not recorded | 6a279e2490966ade919bf7a2 |
| Qwen-plus | 82.84 | Not recorded | 6a279e2490966ade919bf7a2 |
| DeepSeek-R1-Distill-Llama-8B | 78.81 | Not recorded | 6a279e2490966ade919bf7a2 |
| DeepSeek-R1-Distill-Llama-8B | 77.56 | Not recorded | 6a279e2490966ade919bf7a2 |
| DeepSeek-R1 | 65.88 | Not recorded | 6a279e2490966ade919bf7a2 |
| Qwen-7B | 63.24 | Not recorded | 6a279e2490966ade919bf7a2 |

### Qwen-7B

Support F1: 89.9. Result ID: 6a27d67efbc373592c24c2a8.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Fine-Tuning"}. This record does not assert independent reproduction.

### Qwen-Max

Support F1: 89.55. Result ID: 6a27d67efbc373592c24c2a1.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

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

### DeepSeek-V3

Support F1: 89.17. Result ID: 6a27d67efbc373592c24c29e.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77001e209131cea7066a.md>)

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

### Qwen-Max

Support F1: 87.57. Result ID: 6a27d67efbc373592c24c2a5.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"3-shot prompt learning"}. This record does not assert independent reproduction.

### DeepSeek-R1

Support F1: 87.14. Result ID: 6a27d67efbc373592c24c29f.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76a91e209131cea6e3eb.md>)

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

### Qwen-plus

Support F1: 85.92. Result ID: 6a27d67efbc373592c24c2a4.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"3-shot prompt learning"}. This record does not assert independent reproduction.

### DeepSeek-V3

Support F1: 85.33. Result ID: 6a27d67efbc373592c24c2a2.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77001e209131cea7066a.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"3-shot prompt learning"}. This record does not assert independent reproduction.

### Qwen-plus

Support F1: 82.84. Result ID: 6a27d67efbc373592c24c2a0.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

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

### DeepSeek-R1-Distill-Llama-8B

Support F1: 78.81. Result ID: 6a27d67efbc373592c24c2a9.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Fine-Tuning"}. This record does not assert independent reproduction.

### DeepSeek-R1-Distill-Llama-8B

Support F1: 77.56. Result ID: 6a27d67efbc373592c24c2a7.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

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

### DeepSeek-R1

Support F1: 65.88. Result ID: 6a27d67efbc373592c24c2a3.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76a91e209131cea6e3eb.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"3-shot prompt learning"}. This record does not assert independent reproduction.

### Qwen-7B

Support F1: 63.24. Result ID: 6a27d67efbc373592c24c2a6.

[Reporting paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a279e2490966ade919bf7a2.md>)

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

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