# Intrinsic Bias Evaluation on NLI-bias

Source: [Intrinsic Bias Evaluation on NLI-bias](<https://www.sota2.com/research/sota/intrinsic-bias-evaluation-on-nli-bias>)

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

Data updated: 2026-02-26T06:04:47.918Z

Evaluation of intrinsic bias using the StereoSet Score metric on the NLI-bias dataset.

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 | Intrinsic Bias Evaluation |
| Dataset (including any stated split) | NLI-bias |
| Metric | StereoSet Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [StereoSet Score](<https://www.sota2.com/research/sota/intrinsic-bias-evaluation-on-nli-bias.md?metric=StereoSet%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | StereoSet Score | Result date | Report paper |
| --- | --- | --- | --- |
| Debiased-tuning | 54.94 | Not recorded | 68fa738e1e209131cea5d2e0 |
| ProSocialTuning | 53.96 | Not recorded | 68fa738e1e209131cea5d2e0 |
| DEBIASED | 53.2 | Not recorded | 68fa738e1e209131cea5d2e0 |

### Debiased-tuning

StereoSet Score: 54.94. Result ID: 6992ff533a6c31484912f0b4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa738e1e209131cea5d2e0.md>)

Recorded evaluation configuration: {"Mode":"Fine-tuned"}. This record does not assert independent reproduction.

### ProSocialTuning

StereoSet Score: 53.96. Result ID: 6992ff533a6c31484912f0b5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa738e1e209131cea5d2e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738e1e209131cea5d2e0.md>)

[Implementation](<https://github.com/MSU-NLP-CSS/ProSocialTuning>)

Recorded evaluation configuration: {"Mode":"Fine-tuned"}. This record does not assert independent reproduction.

### DEBIASED

StereoSet Score: 53.2. Result ID: 6992ff533a6c31484912f0b3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa738e1e209131cea5d2e0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69751e209131cea22385.md>)

Recorded evaluation configuration: {"Model description":"Debiased model using CDA"}. This record does not assert independent reproduction.

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