# Natural Language Understanding on NLU 21 tasks one-shot

Source: [Natural Language Understanding on NLU 21 tasks one-shot](<https://www.sota2.com/research/sota/natural-language-understanding-on-nlu-21-tasks-one-shot>)

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

Data updated: 2026-02-26T05:35:48.157Z

Evaluation of general Natural Language Understanding capabilities aggregated across 21 distinct NLU tasks, measured using a one-shot prompting strategy.

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 Understanding |
| Dataset (including any stated split) | NLU 21 tasks one-shot |
| Metric | Average Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 8 of 8 |
| 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

- [Average Score](<https://www.sota2.com/research/sota/natural-language-understanding-on-nlu-21-tasks-one-shot.md?metric=Average%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Average Score | Result date | Report paper |
| --- | --- | --- | --- |
| Lion | 61.7 | Not recorded | 68fa6f421e209131cea448e3 |
| Adafactor | 61.3 | Not recorded | 68fa6f421e209131cea448e3 |
| PaLM | 59.4 | Not recorded | 68fa6f421e209131cea448e3 |
| GPT-3 | 58.5 | Not recorded | 68fa6f421e209131cea448e3 |
| Lion | 57.4 | Not recorded | 68fa6f421e209131cea448e3 |
| Adafactor | 56.8 | Not recorded | 68fa6f421e209131cea448e3 |
| Lion | 53.9 | Not recorded | 68fa6f421e209131cea448e3 |
| Adafactor | 53.2 | Not recorded | 68fa6f421e209131cea448e3 |

### Lion

Average Score: 61.7. Result ID: 6990b0ff3a6c314849027aeb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Implementation](<https://github.com/google/automl/tree/master/lion>)

Recorded evaluation configuration: {"Model Size":"7.5B","Optimizer":"Lion","#Tokens":"300B","Evaluation Protocol":"One-shot"}. This record does not assert independent reproduction.

### Adafactor

Average Score: 61.3. Result ID: 6990b0ff3a6c314849027aea.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64471e209131cea03a91.md>)

Recorded evaluation configuration: {"Model Size":"7.5B","Optimizer":"Adafactor","#Tokens":"300B","Evaluation Protocol":"One-shot"}. This record does not assert independent reproduction.

### PaLM

Average Score: 59.4. Result ID: 6990b0ff3a6c314849027aed.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d931e209131cea3abf2.md>)

Recorded evaluation configuration: {"Model Size":"8B","Optimizer":"not specified","#Tokens":"780B","Evaluation Protocol":"One-shot"}. This record does not assert independent reproduction.

### GPT-3

Average Score: 58.5. Result ID: 6990b0ff3a6c314849027aec.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa680c1e209131cea19d53.md>)

[Implementation](<https://github.com/openai/gpt-3/blob/master/overlap_frequency.md>)

Recorded evaluation configuration: {"Model Size":"6.7B","Optimizer":"AdamW (implied)","#Tokens":"300B","Evaluation Protocol":"One-shot"}. This record does not assert independent reproduction.

### Lion

Average Score: 57.4. Result ID: 6990b0ff3a6c314849027ae9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Implementation](<https://github.com/google/automl/tree/master/lion>)

Recorded evaluation configuration: {"Model Size":"2.1B","Optimizer":"Lion","#Tokens":"300B","Evaluation Protocol":"One-shot"}. This record does not assert independent reproduction.

### Adafactor

Average Score: 56.8. Result ID: 6990b0ff3a6c314849027ae8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64471e209131cea03a91.md>)

Recorded evaluation configuration: {"Model Size":"2.1B","Optimizer":"Adafactor","#Tokens":"300B","Evaluation Protocol":"One-shot"}. This record does not assert independent reproduction.

### Lion

Average Score: 53.9. Result ID: 6990b0ff3a6c314849027ae7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Implementation](<https://github.com/google/automl/tree/master/lion>)

Recorded evaluation configuration: {"Model Size":"1.1B","Optimizer":"Lion","#Tokens":"300B","Evaluation Protocol":"One-shot"}. This record does not assert independent reproduction.

### Adafactor

Average Score: 53.2. Result ID: 6990b0ff3a6c314849027ae6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6f421e209131cea448e3.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa64471e209131cea03a91.md>)

Recorded evaluation configuration: {"Model Size":"1.1B","Optimizer":"Adafactor","#Tokens":"300B","Evaluation Protocol":"One-shot"}. This record does not assert independent reproduction.

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