# Natural Language Understanding on SuperGLUE 1,000 examples (test)

Source: [Natural Language Understanding on SuperGLUE 1,000 examples (test)](<https://www.sota2.com/research/sota/natural-language-understanding-on-superglue-1-000-examples-test>)

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

Data updated: 2026-02-26T05:37:53.508Z

Evaluation of general language understanding capabilities using a specific 1,000-example subset of the SuperGLUE benchmark suite, evaluated on the test split.

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) | SuperGLUE 1,000 examples (test) |
| Metric | BoolQ |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 10 of 10 |
| 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

- [BoolQ](<https://www.sota2.com/research/sota/natural-language-understanding-on-superglue-1-000-examples-test.md?metric=BoolQ>): Higher is better; unit: PERCENT.

- [RTE](<https://www.sota2.com/research/sota/natural-language-understanding-on-superglue-1-000-examples-test.md?metric=RTE>): Higher is better; unit: PERCENT.

- [WiC](<https://www.sota2.com/research/sota/natural-language-understanding-on-superglue-1-000-examples-test.md?metric=WiC>): Higher is better; unit: PERCENT.

- [MultiRC](<https://www.sota2.com/research/sota/natural-language-understanding-on-superglue-1-000-examples-test.md?metric=MultiRC>): Higher is better; unit: PERCENT.

- [SST-2](<https://www.sota2.com/research/sota/natural-language-understanding-on-superglue-1-000-examples-test.md?metric=SST-2>): Higher is better; unit: PERCENT.

- [COPA](<https://www.sota2.com/research/sota/natural-language-understanding-on-superglue-1-000-examples-test.md?metric=COPA>): Higher is better; unit: PERCENT.

- [Average Score](<https://www.sota2.com/research/sota/natural-language-understanding-on-superglue-1-000-examples-test.md?metric=Average%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | BoolQ | Result date | Report paper |
| --- | --- | --- | --- |
| FT | 86.7 | Not recorded | 68fa725a1e209131cea55e46 |
| S-MeZO | 85.3 | Not recorded | 68fa725a1e209131cea55e46 |
| LoRA | 84.8 | Not recorded | 68fa725a1e209131cea55e46 |
| R-MeZO | 84 | Not recorded | 68fa725a1e209131cea55e46 |
| MeZO - LoRA | 83.5 | Not recorded | 68fa725a1e209131cea55e46 |
| MeZO | 81.6 | Not recorded | 68fa725a1e209131cea55e46 |
| ICL | 76.7 | Not recorded | 68fa725a1e209131cea55e46 |
| Zero-Shot | 69.3 | Not recorded | 68fa725a1e209131cea55e46 |
| ICL | 67.4 | Not recorded | 68fa725a1e209131cea55e46 |
| Zero-Shot | 65.1 | Not recorded | 68fa725a1e209131cea55e46 |

### FT

BoolQ: 86.7. Result ID: 699a9efd5698665ec42adbd2.

[Reporting paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.md>)

Recorded evaluation configuration: {"Model":"Mistral-7b","Optimization":"Full-parameter fine-tuning with Adam"}. This record does not assert independent reproduction.

### S-MeZO

BoolQ: 85.3. Result ID: 699a9efd5698665ec42adbd6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.md>)

Recorded evaluation configuration: {"Model":"Mistral-7b","Variant":"Sparse MeZO"}. This record does not assert independent reproduction.

### LoRA

BoolQ: 84.8. Result ID: 699a9efd5698665ec42adbd1.

[Reporting paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.md>)

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

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

Recorded evaluation configuration: {"Model":"Mistral-7b"}. This record does not assert independent reproduction.

### R-MeZO

BoolQ: 84. Result ID: 699a9efd5698665ec42adbd5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.md>)

Recorded evaluation configuration: {"Model":"Mistral-7b","Variant":"MeZO with Random Mask"}. This record does not assert independent reproduction.

### MeZO - LoRA

BoolQ: 83.5. Result ID: 699a9efd5698665ec42adbd4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.md>)

Recorded evaluation configuration: {"Model":"Mistral-7b"}. This record does not assert independent reproduction.

### MeZO

BoolQ: 81.6. Result ID: 699a9efd5698665ec42adbd3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71d41e209131cea5338c.md>)

Recorded evaluation configuration: {"Model":"Mistral-7b"}. This record does not assert independent reproduction.

### ICL

BoolQ: 76.7. Result ID: 699a9efd5698665ec42adbd0.

[Reporting paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.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":"Mistral-7b","Strategy":"In-Context Learning"}. This record does not assert independent reproduction.

### Zero-Shot

BoolQ: 69.3. Result ID: 699a9efd5698665ec42adbcf.

[Reporting paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.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":"Mistral-7b"}. This record does not assert independent reproduction.

### ICL

BoolQ: 67.4. Result ID: 699a9efd5698665ec42adbbc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.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":"LLaMA-7b","Protocol":"In-Context Learning"}. This record does not assert independent reproduction.

### Zero-Shot

BoolQ: 65.1. Result ID: 699a9efd5698665ec42adbbb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa725a1e209131cea55e46.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":"LLaMA-7b"}. This record does not assert independent reproduction.

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