# Natural Language Explanation Generation on SBIC (test)

Source: [Natural Language Explanation Generation on SBIC (test)](<https://www.sota2.com/research/sota/natural-language-explanation-generation-on-sbic-test>)

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

Data updated: 2026-02-26T05:29:55.283Z

Evaluation of model performance on generating natural language explanations using the test split of the SBIC dataset, measured using metrics like Accuracy (Acc.) and nBERTs.

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 | Natural Language Explanation Generation |
| Dataset (including any stated split) | SBIC (test) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher 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

- [Accuracy](<https://www.sota2.com/research/sota/natural-language-explanation-generation-on-sbic-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [nBERTs Score](<https://www.sota2.com/research/sota/natural-language-explanation-generation-on-sbic-test.md?metric=nBERTs%20Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| SPARSEFIT (Att.Q+LN) | 66.99 | Not recorded | 68fa74701e209131cea61c52 |
| LoRA (Att.QV, Rank=128) | 63.57 | Not recorded | 68fa74701e209131cea61c52 |
| LoRA (Att.KQVO, Rank=4) | 62.66 | Not recorded | 68fa74701e209131cea61c52 |
| AdaLoRA (1.15%) | 61.8 | Not recorded | 68fa74701e209131cea61c52 |
| (IA)³ | 48.08 | Not recorded | 68fa74701e209131cea61c52 |
| AdaLoRA (4.48%) | 0.18 | Not recorded | 68fa74701e209131cea61c52 |

### SPARSEFIT (Att.Q+LN)

Accuracy: 66.99. Result ID: 69935ccd3a6c3148491568bb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74701e209131cea61c52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74701e209131cea61c52.md>)

Recorded evaluation configuration: {"Trainable Parameters":"6.84%","FLOPS":"2.37e14","Base Model":"T5-large"}. This record does not assert independent reproduction.

### LoRA (Att.QV, Rank=128)

Accuracy: 63.57. Result ID: 69935ccd3a6c3148491568be.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74701e209131cea61c52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b121e209131cea2bf59.md>)

[Implementation](<https://github.com/microsoft/LoRA>)

Recorded evaluation configuration: {"Trainable Parameters":"4.86%","FLOPS":"2.88e14","Base Model":"T5-large"}. This record does not assert independent reproduction.

### LoRA (Att.KQVO, Rank=4)

Accuracy: 62.66. Result ID: 69935ccd3a6c3148491568bf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74701e209131cea61c52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6b121e209131cea2bf59.md>)

[Implementation](<https://github.com/microsoft/LoRA>)

Recorded evaluation configuration: {"Trainable Parameters":"0.32%","FLOPS":"2.75e14","Base Model":"T5-large"}. This record does not assert independent reproduction.

### AdaLoRA (1.15%)

Accuracy: 61.8. Result ID: 69935ccd3a6c3148491568bd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74701e209131cea61c52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71861e209131cea51c24.md>)

Recorded evaluation configuration: {"Trainable Parameters":"1.15%","FLOPS":"1.48e15","Base Model":"T5-large"}. This record does not assert independent reproduction.

### (IA)³

Accuracy: 48.08. Result ID: 69935ccd3a6c3148491568c0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74701e209131cea61c52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d491e209131cea38c7f.md>)

[Implementation](<https://github.com/r-three/t-few>)

Recorded evaluation configuration: {"Trainable Parameters":"0.07%","FLOPS":"2.74e14","Base Model":"T5-large"}. This record does not assert independent reproduction.

### AdaLoRA (4.48%)

Accuracy: 0.18. Result ID: 69935ccd3a6c3148491568bc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa74701e209131cea61c52.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71861e209131cea51c24.md>)

Recorded evaluation configuration: {"Trainable Parameters":"4.48%","FLOPS":"2.87e14","Base Model":"T5-large"}. This record does not assert independent reproduction.

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