# Zero-shot Evaluation on Downstream NLU tasks 1.0 (eval)

Source: [Zero-shot Evaluation on Downstream NLU tasks 1.0 (eval)](<https://www.sota2.com/research/sota/zero-shot-evaluation-on-downstream-nlu-tasks-1-0-eval>)

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

Data updated: 2026-02-26T05:21:00.555Z

Evaluation of zero-shot performance across various downstream Natural Language Understanding tasks included in the 'Downstream NLU tasks 1.0' benchmark suite, measured by the 0-shot Score.

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 | Zero-shot Evaluation |
| Dataset (including any stated split) | Downstream NLU tasks 1.0 (eval) |
| Metric | 0-shot Score |
| 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

- [0-shot Score](<https://www.sota2.com/research/sota/zero-shot-evaluation-on-downstream-nlu-tasks-1-0-eval.md?metric=0-shot%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | 0-shot Score | Result date | Report paper |
| --- | --- | --- | --- |
| Random Sampling KD (Ours 12+) | 57.9 | Not recorded | 68fa7a291e209131cea816e3 |
| Random Sampling KD (Ours 12) | 57.5 | Not recorded | 68fa7a291e209131cea816e3 |
| FullKD | 57.5 | Not recorded | 68fa7a291e209131cea816e3 |
| Top-K 50 | 57.1 | Not recorded | 68fa7a291e209131cea816e3 |
| Top-K 12 | 56.6 | Not recorded | 68fa7a291e209131cea816e3 |
| CE | 55.6 | Not recorded | 68fa7a291e209131cea816e3 |

### Random Sampling KD (Ours 12+)

0-shot Score: 57.9. Result ID: 698b6f707a967eddb76e8e7e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a291e209131cea816e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a291e209131cea816e3.md>)

Recorded evaluation configuration: {"Unique Tokens":12,"Adaptive Weighting":true}. This record does not assert independent reproduction.

### Random Sampling KD (Ours 12)

0-shot Score: 57.5. Result ID: 698b6f707a967eddb76e8e7d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a291e209131cea816e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a291e209131cea816e3.md>)

Recorded evaluation configuration: {"Unique Tokens":12}. This record does not assert independent reproduction.

### FullKD

0-shot Score: 57.5. Result ID: 698b6f707a967eddb76e8e7f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a291e209131cea816e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74551e209131cea61267.md>)

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

### Top-K 50

0-shot Score: 57.1. Result ID: 698b6f707a967eddb76e8e7c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a291e209131cea816e3.md>)

Recorded evaluation configuration: {"K":50}. This record does not assert independent reproduction.

### Top-K 12

0-shot Score: 56.6. Result ID: 698b6f707a967eddb76e8e7b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a291e209131cea816e3.md>)

Recorded evaluation configuration: {"K":12}. This record does not assert independent reproduction.

### CE

0-shot Score: 55.6. Result ID: 698b6f707a967eddb76e8e7a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7a291e209131cea816e3.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.