# Multi-step Reasoning on Bamboogle auto-eval (test)

Source: [Multi-step Reasoning on Bamboogle auto-eval (test)](<https://www.sota2.com/research/sota/multi-step-reasoning-on-bamboogle-auto-eval-test>)

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

Data updated: 2026-02-26T05:51:51.390Z

Evaluation of model performance on multi-step reasoning tasks using the test split of the Bamboogle auto-eval dataset, reporting mean accuracy and standard deviation.

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 | Multi-step Reasoning |
| Dataset (including any stated split) | Bamboogle auto-eval (test) |
| Metric | Mean Accuracy |
| 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

- [Mean Accuracy](<https://www.sota2.com/research/sota/multi-step-reasoning-on-bamboogle-auto-eval-test.md?metric=Mean%20Accuracy>): Higher is better; unit: PERCENT.

- [Standard Deviation](<https://www.sota2.com/research/sota/multi-step-reasoning-on-bamboogle-auto-eval-test.md?metric=Standard%20Deviation>): Lower is better; unit: SCALAR.

## Recorded results

| Method | Mean Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Self-improvement, 2nd gen | 76.1 | Not recorded | 68fa71771e209131cea516a5 |
| Self-improvement, 1st gen | 74 | Not recorded | 68fa71771e209131cea516a5 |
| Pilot, human filtered | 71.5 | Not recorded | 68fa71771e209131cea516a5 |
| Pre-trained | 70.3 | Not recorded | 68fa71771e209131cea516a5 |
| Self-improvement, 2nd gen | 69.7 | Not recorded | 68fa71771e209131cea516a5 |
| Self-improvement, 2nd gen | 65.9 | Not recorded | 68fa71771e209131cea516a5 |
| Self-improvement, 1st gen | 61.9 | Not recorded | 68fa71771e209131cea516a5 |
| Pilot, human filtered | 56.6 | Not recorded | 68fa71771e209131cea516a5 |
| Self-improvement, 1st gen | 54.4 | Not recorded | 68fa71771e209131cea516a5 |
| Pilot, human filtered | 44.7 | Not recorded | 68fa71771e209131cea516a5 |

### Self-improvement, 2nd gen

Mean Accuracy: 76.1. Result ID: 6991d0063a6c3148490abe15.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

Recorded evaluation configuration: {"Model Size":"L"}. This record does not assert independent reproduction.

### Self-improvement, 1st gen

Mean Accuracy: 74. Result ID: 6991d0063a6c3148490abe12.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

Recorded evaluation configuration: {"Model Size":"L"}. This record does not assert independent reproduction.

### Pilot, human filtered

Mean Accuracy: 71.5. Result ID: 6991d0063a6c3148490abe0f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

Recorded evaluation configuration: {"Model Size":"L"}. This record does not assert independent reproduction.

### Pre-trained

Mean Accuracy: 70.3. Result ID: 6991d0063a6c3148490abe0c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa708e1e209131cea4bc0a.md>)

Recorded evaluation configuration: {"Model Size":"L"}. This record does not assert independent reproduction.

### Self-improvement, 2nd gen

Mean Accuracy: 69.7. Result ID: 6991d0063a6c3148490abe14.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

Recorded evaluation configuration: {"Model Size":"S"}. This record does not assert independent reproduction.

### Self-improvement, 2nd gen

Mean Accuracy: 65.9. Result ID: 6991d0063a6c3148490abe13.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

Recorded evaluation configuration: {"Model Size":"XS"}. This record does not assert independent reproduction.

### Self-improvement, 1st gen

Mean Accuracy: 61.9. Result ID: 6991d0063a6c3148490abe11.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

Recorded evaluation configuration: {"Model Size":"S"}. This record does not assert independent reproduction.

### Pilot, human filtered

Mean Accuracy: 56.6. Result ID: 6991d0063a6c3148490abe0e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

Recorded evaluation configuration: {"Model Size":"S"}. This record does not assert independent reproduction.

### Self-improvement, 1st gen

Mean Accuracy: 54.4. Result ID: 6991d0063a6c3148490abe10.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

Recorded evaluation configuration: {"Model Size":"XS"}. This record does not assert independent reproduction.

### Pilot, human filtered

Mean Accuracy: 44.7. Result ID: 6991d0063a6c3148490abe0d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71771e209131cea516a5.md>)

Recorded evaluation configuration: {"Model Size":"XS"}. This record does not assert independent reproduction.

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