# Harmfulness Assessment on Bi-directional Anchoring evaluation set

Source: [Harmfulness Assessment on Bi-directional Anchoring evaluation set](<https://www.sota2.com/research/sota/harmfulness-assessment-on-bi-directional-anchoring-evaluation-set>)

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

Data updated: 2026-02-26T05:46:23.258Z

Evaluation of model performance on the Bi-directional Anchoring evaluation set using the Harmfulness Assessment task.

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 | Harmfulness Assessment |
| Dataset (including any stated split) | Bi-directional Anchoring evaluation set |
| Metric | Harmful Score |
| Unit | SCALAR |
| Direction | Lower 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

- [Harmful Score](<https://www.sota2.com/research/sota/harmfulness-assessment-on-bi-directional-anchoring-evaluation-set.md?metric=Harmful%20Score>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Harmful Score | Result date | Report paper |
| --- | --- | --- | --- |
| Vaccine-SFT | 28.5 | Not recorded | 68fa756a1e209131cea67bac |
| Vaccine-SFT | 30.3 | Not recorded | 68fa756a1e209131cea67bac |
| SFT | 33.9 | Not recorded | 68fa756a1e209131cea67bac |
| Lisa | 33.9 | Not recorded | 68fa756a1e209131cea67bac |
| Lisa | 33.9 | Not recorded | 68fa756a1e209131cea67bac |
| SFT | 34.6 | Not recorded | 68fa756a1e209131cea67bac |

### Vaccine-SFT

Harmful Score: 28.5. Result ID: 6993a0d43a6c314849176217.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756a1e209131cea67bac.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75d01e209131cea69fa9.md>)

[Implementation](<https://github.com/git-disl/Vaccine>)

Recorded evaluation configuration: {"Training Stage":"before fine-tuning"}. This record does not assert independent reproduction.

### Vaccine-SFT

Harmful Score: 30.3. Result ID: 6993a0d43a6c314849176218.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756a1e209131cea67bac.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75d01e209131cea69fa9.md>)

[Implementation](<https://github.com/git-disl/Vaccine>)

Recorded evaluation configuration: {"Training Stage":"after fine-tuning"}. This record does not assert independent reproduction.

### SFT

Harmful Score: 33.9. Result ID: 6993a0d43a6c314849176215.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756a1e209131cea67bac.md>)

Recorded evaluation configuration: {"Training Stage":"before fine-tuning"}. This record does not assert independent reproduction.

### Lisa

Harmful Score: 33.9. Result ID: 6993a0d43a6c314849176219.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756a1e209131cea67bac.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa756a1e209131cea67bac.md>)

[Implementation](<https://github.com/git-disl/Lisa>)

Recorded evaluation configuration: {"Training Stage":"before fine-tuning"}. This record does not assert independent reproduction.

### Lisa

Harmful Score: 33.9. Result ID: 6993a0d43a6c31484917621a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756a1e209131cea67bac.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa756a1e209131cea67bac.md>)

[Implementation](<https://github.com/git-disl/Lisa>)

Recorded evaluation configuration: {"Training Stage":"after fine-tuning"}. This record does not assert independent reproduction.

### SFT

Harmful Score: 34.6. Result ID: 6993a0d43a6c314849176216.

[Reporting paper](<https://www.sota2.com/research/paper/68fa756a1e209131cea67bac.md>)

Recorded evaluation configuration: {"Training Stage":"after fine-tuning"}. This record does not assert independent reproduction.

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