# Step-level quality assessment on PRMBench

Source: [Step-level quality assessment on PRMBench](<https://www.sota2.com/research/sota/step-level-quality-assessment-on-prmbench>)

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

Data updated: 2026-04-14T09:53:42.660Z

Evaluation of model performance on step-level quality assessment using the PRMBench dataset, focusing on metrics like Simplicity, Soundness, and Sensitivity.

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 | Step-level quality assessment |
| Dataset (including any stated split) | PRMBench |
| Metric | Simplicity |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| 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

- [Simplicity](<https://www.sota2.com/research/sota/step-level-quality-assessment-on-prmbench.md?metric=Simplicity>): Higher is better; unit: PERCENT.

- [Soundness](<https://www.sota2.com/research/sota/step-level-quality-assessment-on-prmbench.md?metric=Soundness>): Higher is better; unit: PERCENT.

- [Sensitivity](<https://www.sota2.com/research/sota/step-level-quality-assessment-on-prmbench.md?metric=Sensitivity>): Higher is better; unit: PERCENT.

- [Total Score](<https://www.sota2.com/research/sota/step-level-quality-assessment-on-prmbench.md?metric=Total%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Simplicity | Result date | Report paper |
| --- | --- | --- | --- |
| CPMI\_Merge | 54.65 | Not recorded | 69ddca1a7d0172140ec79a51 |
| CPMI | 52.56 | Not recorded | 69ddca1a7d0172140ec79a51 |
| PAV | 47.16 | Not recorded | 69ddca1a7d0172140ec79a51 |
| MC | 40.88 | Not recorded | 69ddca1a7d0172140ec79a51 |
| Rand\_Merge | 27.65 | Not recorded | 69ddca1a7d0172140ec79a51 |

### CPMI\_Merge

Simplicity: 54.65. Result ID: 69de0e9e47f34ce924c5f069.

[Reporting paper](<https://www.sota2.com/research/paper/69ddca1a7d0172140ec79a51.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ddca1a7d0172140ec79a51.md>)

Recorded evaluation configuration: {"Reward Type":"CPMI\_Merge"}. This record does not assert independent reproduction.

### CPMI

Simplicity: 52.56. Result ID: 69de0e9e47f34ce924c5f068.

[Reporting paper](<https://www.sota2.com/research/paper/69ddca1a7d0172140ec79a51.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69ddca1a7d0172140ec79a51.md>)

Recorded evaluation configuration: {"Reward Type":"CPMI"}. This record does not assert independent reproduction.

### PAV

Simplicity: 47.16. Result ID: 69de0e9e47f34ce924c5f067.

[Reporting paper](<https://www.sota2.com/research/paper/69ddca1a7d0172140ec79a51.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75201e209131cea65f64.md>)

Recorded evaluation configuration: {"Reward Type":"PAV"}. This record does not assert independent reproduction.

### MC

Simplicity: 40.88. Result ID: 69de0e9e47f34ce924c5f066.

[Reporting paper](<https://www.sota2.com/research/paper/69ddca1a7d0172140ec79a51.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72381e209131cea550cf.md>)

Recorded evaluation configuration: {"Reward Type":"MC"}. This record does not assert independent reproduction.

### Rand\_Merge

Simplicity: 27.65. Result ID: 69de0e9e47f34ce924c5f06a.

[Reporting paper](<https://www.sota2.com/research/paper/69ddca1a7d0172140ec79a51.md>)

Recorded evaluation configuration: {"Reward Type":"Rand\_Merge"}. This record does not assert independent reproduction.

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