# Rationalization on HateXplain Synthetic Skew (test)

Source: [Rationalization on HateXplain Synthetic Skew (test)](<https://www.sota2.com/research/sota/rationalization-on-hatexplain-synthetic-skew-test>)

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

Data updated: 2026-02-26T05:46:10.624Z

Evaluation of Rationalization performance on the synthetic skewed test split of the HateXplain dataset, using HI-F1, R, and S metrics.

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 | Rationalization |
| Dataset (including any stated split) | HateXplain Synthetic Skew (test) |
| Metric | HI-F1 |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 7 of 7 |
| 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

- [HI-F1](<https://www.sota2.com/research/sota/rationalization-on-hatexplain-synthetic-skew-test.md?metric=HI-F1>): Higher is better; unit: SCALAR.

- [R](<https://www.sota2.com/research/sota/rationalization-on-hatexplain-synthetic-skew-test.md?metric=R>): Higher is better; unit: SCALAR.

- [S](<https://www.sota2.com/research/sota/rationalization-on-hatexplain-synthetic-skew-test.md?metric=S>): Higher is better; unit: SCALAR.

## Recorded results

| Method | HI-F1 | Result date | Report paper |
| --- | --- | --- | --- |
| GenSPP | 0.4281 | Not recorded | 68fa78dc1e209131cea7ad91 |
| GenSPP | 0.3989 | Not recorded | 68fa78dc1e209131cea7ad91 |
| GenSPP | 0.3552 | Not recorded | 68fa78dc1e209131cea7ad91 |
| G-RAT | 0.3433 | Not recorded | 68fa78dc1e209131cea7ad91 |
| MGR | 0.1445 | Not recorded | 68fa78dc1e209131cea7ad91 |
| MCD | 0.1388 | Not recorded | 68fa78dc1e209131cea7ad91 |
| FR | 0.0722 | Not recorded | 68fa78dc1e209131cea7ad91 |

### GenSPP

HI-F1: 0.4281. Result ID: 6995a8913badfaffe6a6986a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78dc1e209131cea7ad91.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78dc1e209131cea7ad91.md>)

[Implementation](<https://github.com/federicoruggeri/gen-spp>)

Recorded evaluation configuration: {"G":150,"Initialization strategy":"skewed individual"}. This record does not assert independent reproduction.

### GenSPP

HI-F1: 0.3989. Result ID: 6995a8913badfaffe6a69869.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78dc1e209131cea7ad91.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78dc1e209131cea7ad91.md>)

[Implementation](<https://github.com/federicoruggeri/gen-spp>)

Recorded evaluation configuration: {"G":100,"Initialization strategy":"skewed individual"}. This record does not assert independent reproduction.

### GenSPP

HI-F1: 0.3552. Result ID: 6995a8913badfaffe6a6986b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78dc1e209131cea7ad91.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78dc1e209131cea7ad91.md>)

[Implementation](<https://github.com/federicoruggeri/gen-spp>)

Recorded evaluation configuration: {"G":100,"Initialization strategy":"Gaussian noise variants"}. This record does not assert independent reproduction.

### G-RAT

HI-F1: 0.3433. Result ID: 6995a8913badfaffe6a69868.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78dc1e209131cea7ad91.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### MGR

HI-F1: 0.1445. Result ID: 6995a8913badfaffe6a69866.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78dc1e209131cea7ad91.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa701b1e209131cea4909a.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### MCD

HI-F1: 0.1388. Result ID: 6995a8913badfaffe6a69867.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78dc1e209131cea7ad91.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71141e209131cea4f213.md>)

[Implementation](<https://github.com/jugechengzi/Rationalization-MCD>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### FR

HI-F1: 0.0722. Result ID: 6995a8913badfaffe6a69865.

[Reporting paper](<https://www.sota2.com/research/paper/68fa78dc1e209131cea7ad91.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d751e209131cea3a111.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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