# Toxicity Classification on CivilComments-WILDS

Source: [Toxicity Classification on CivilComments-WILDS](<https://www.sota2.com/research/sota/toxicity-classification-on-civilcomments-wilds>)

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

Data updated: 2026-04-21T08:31:51.281Z

Evaluation of Toxicity Classification performance on the CivilComments WILDS dataset, reporting group fairness metrics such as Worst-Group Accuracy.

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 | Toxicity Classification |
| Dataset (including any stated split) | CivilComments WILDS |
| Metric | Worst-Group Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 11 of 11 |
| 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

- [Worst-Group Accuracy](<https://www.sota2.com/research/sota/toxicity-classification-on-civilcomments-wilds.md?metric=Worst-Group%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Worst-Group Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Fish | 75.3 | Not recorded | 68fa76241e209131cea6b73b |
| IRMX (PAIR) | 74.2 | Not recorded | 68fa76241e209131cea6b73b |
| IRMX | 73.4 | Not recorded | 68fa76241e209131cea6b73b |
| LISA | 72.9 | Not recorded | 68fa76241e209131cea6b73b |
| DFR | 72.5 | Not recorded | 68fa76241e209131cea6b73b |
| Group DRO | 70 | Not recorded | 68fa76241e209131cea6b73b |
| IRM | 66.3 | Not recorded | 68fa76241e209131cea6b73b |
| CORAL | 65.6 | Not recorded | 68fa76241e209131cea6b73b |
| Shared-Base | 58.1 | Not recorded | 68fa76241e209131cea6b73b |
| Shared-Base + HyRe | 58.1 | Not recorded | 68fa76241e209131cea6b73b |
| ERM | 56 | Not recorded | 68fa76241e209131cea6b73b |

### Fish

Worst-Group Accuracy: 75.3. Result ID: 69e735f147f34ce924cd6cbe.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a561e209131cea2785c.md>)

[Implementation](<https://github.com/YugeTen/fish>)

Recorded evaluation configuration: {"DL":"O"}. This record does not assert independent reproduction.

### IRMX (PAIR)

Worst-Group Accuracy: 74.2. Result ID: 69e735f147f34ce924cd6cba.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

Recorded evaluation configuration: {"DL":"O"}. This record does not assert independent reproduction.

### IRMX

Worst-Group Accuracy: 73.4. Result ID: 69e735f147f34ce924cd6cb9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

Recorded evaluation configuration: {"DL":"O"}. This record does not assert independent reproduction.

### LISA

Worst-Group Accuracy: 72.9. Result ID: 69e735f147f34ce924cd6cbf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6cd91e209131cea36415.md>)

Recorded evaluation configuration: {"DL":"O"}. This record does not assert independent reproduction.

### DFR

Worst-Group Accuracy: 72.5. Result ID: 69e735f147f34ce924cd6cbd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

Recorded evaluation configuration: {"DL":"O"}. This record does not assert independent reproduction.

### Group DRO

Worst-Group Accuracy: 70. Result ID: 69e735f147f34ce924cd6cbc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67561e209131cea159c7.md>)

Recorded evaluation configuration: {"DL":"O"}. This record does not assert independent reproduction.

### IRM

Worst-Group Accuracy: 66.3. Result ID: 69e735f147f34ce924cd6cb8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa673c1e209131cea15133.md>)

[Implementation](<https://github.com/facebookresearch/InvariantRiskMinimization>)

Recorded evaluation configuration: {"DL":"O"}. This record does not assert independent reproduction.

### CORAL

Worst-Group Accuracy: 65.6. Result ID: 69e735f147f34ce924cd6cbb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62fe1e209131ce9fc1c4.md>)

[Implementation](<https://github.com/VisionLearningGroup/CORAL>)

Recorded evaluation configuration: {"DL":"O"}. This record does not assert independent reproduction.

### Shared-Base

Worst-Group Accuracy: 58.1. Result ID: 69e735f147f34ce924cd6cc1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

Recorded evaluation configuration: {"DL":"X"}. This record does not assert independent reproduction.

### Shared-Base + HyRe

Worst-Group Accuracy: 58.1. Result ID: 69e735f147f34ce924cd6cc2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

Recorded evaluation configuration: {"DL":"X"}. This record does not assert independent reproduction.

### ERM

Worst-Group Accuracy: 56. Result ID: 69e735f147f34ce924cd6cc0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa76241e209131cea6b73b.md>)

Recorded evaluation configuration: {"DL":"X"}. This record does not assert independent reproduction.

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