# Downstream accuracy extrapolation on SocialIQA

Source: [Downstream accuracy extrapolation on SocialIQA](<https://www.sota2.com/research/sota/downstream-accuracy-extrapolation-on-socialiqa>)

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

Data updated: 2026-05-19T13:33:12.571Z

Evaluation of model performance on the SocialIQA dataset specifically for downstream accuracy extrapolation, typically measured using Root Mean Squared Error (RMSE).

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 | Downstream accuracy extrapolation |
| Dataset (including any stated split) | SocialIQA |
| Metric | RMSE |
| Unit | SCALAR |
| Direction | Lower 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

- [RMSE](<https://www.sota2.com/research/sota/downstream-accuracy-extrapolation-on-socialiqa.md?metric=RMSE>): Lower is better; unit: SCALAR.

## Recorded results

| Method | RMSE | Result date | Report paper |
| --- | --- | --- | --- |
| CE Loss | 0.011 | Not recorded | 6a0bef1adbc836b87d82cb4e |
| Proxy | 0.016 | Not recorded | 6a0bef1adbc836b87d82cb4e |
| Compute | 0.025 | Not recorded | 6a0bef1adbc836b87d82cb4e |

### CE Loss

RMSE: 0.011. Result ID: 6a0c65affbc373592c042f2a.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bef1adbc836b87d82cb4e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73b21e209131cea5df64.md>)

Recorded evaluation configuration: {"Predictor":"CE Loss"}. This record does not assert independent reproduction.

### Proxy

RMSE: 0.016. Result ID: 6a0c65affbc373592c042f29.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bef1adbc836b87d82cb4e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a0bef1adbc836b87d82cb4e.md>)

[Implementation](<https://github.com/McGill-NLP/proxy-metrics>)

Recorded evaluation configuration: {"Predictor":"Proxy"}. This record does not assert independent reproduction.

### Compute

RMSE: 0.025. Result ID: 6a0c65affbc373592c042f2b.

[Reporting paper](<https://www.sota2.com/research/paper/6a0bef1adbc836b87d82cb4e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71d31e209131cea53301.md>)

Recorded evaluation configuration: {"Predictor":"Compute"}. This record does not assert independent reproduction.

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