# Attribution Faithfulness on SNLI

Source: [Attribution Faithfulness on SNLI](<https://www.sota2.com/research/sota/attribution-faithfulness-on-snli>)

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

Data updated: 2026-02-26T05:45:11.246Z

Evaluation of model attribution faithfulness on the Stanford Natural Language Inference (SNLI) dataset, measured using the 'Faithfulness score' metric.

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 | Attribution Faithfulness |
| Dataset (including any stated split) | SNLI |
| Metric | Faithfulness Score |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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

- [Faithfulness Score](<https://www.sota2.com/research/sota/attribution-faithfulness-on-snli.md?metric=Faithfulness%20Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Faithfulness Score | Result date | Report paper |
| --- | --- | --- | --- |
| Tulu-2 7B | 0.457 | Not recorded | 6937ad576cfd488761a4ec96 |
| Tulu-2 13B | 0.299 | Not recorded | 6937ad576cfd488761a4ec96 |
| Tulu-2 7B | 0.199 | Not recorded | 6937ad576cfd488761a4ec96 |
| Tulu-2 13B | 0.182 | Not recorded | 6937ad576cfd488761a4ec96 |
| Tulu-2 13B | 0.177 | Not recorded | 6937ad576cfd488761a4ec96 |
| Tulu-2 7B | 0.161 | Not recorded | 6937ad576cfd488761a4ec96 |

### Tulu-2 7B

Faithfulness Score: 0.457. Result ID: 6996c82b5698665ec413d37d.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

Recorded evaluation configuration: {"Training Setup":"w/ Explanations","Word Constraint":"One-word"}. This record does not assert independent reproduction.

### Tulu-2 13B

Faithfulness Score: 0.299. Result ID: 6996c82b5698665ec413d380.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

Recorded evaluation configuration: {"Training Setup":"w/ Explanations","Word Constraint":"One-word"}. This record does not assert independent reproduction.

### Tulu-2 7B

Faithfulness Score: 0.199. Result ID: 6996c82b5698665ec413d37b.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

Recorded evaluation configuration: {"Training Setup":"No-Training","Word Constraint":"One-word"}. This record does not assert independent reproduction.

### Tulu-2 13B

Faithfulness Score: 0.182. Result ID: 6996c82b5698665ec413d37f.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

Recorded evaluation configuration: {"Training Setup":"w/ Predictions","Word Constraint":"One-word"}. This record does not assert independent reproduction.

### Tulu-2 13B

Faithfulness Score: 0.177. Result ID: 6996c82b5698665ec413d37e.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

Recorded evaluation configuration: {"Training Setup":"No-Training","Word Constraint":"One-word"}. This record does not assert independent reproduction.

### Tulu-2 7B

Faithfulness Score: 0.161. Result ID: 6996c82b5698665ec413d37c.

[Reporting paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6937ad576cfd488761a4ec96.md>)

Recorded evaluation configuration: {"Training Setup":"w/ Predictions","Word Constraint":"One-word"}. This record does not assert independent reproduction.

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