# Role-Play Faithfulness on PRP - Characters with persona statements at scale (test)

Source: [Role-Play Faithfulness on PRP - Characters with persona statements at scale (test)](<https://www.sota2.com/research/sota/role-play-faithfulness-on-prp-characters-with-persona-statements-at-scale-test>)

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

Data updated: 2026-02-26T06:04:32.178Z

Evaluation of Role-Play Faithfulness performance on the test split of the PRP dataset, measuring adherence to specific character personas (e.g., Spartacus, Cleopatra) using APC metrics.

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 | Role-Play Faithfulness |
| Dataset (including any stated split) | PRP - Characters with persona statements at scale (test) |
| Metric | Spartacus (APC) |
| Unit | SCALAR |
| 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

- [Spartacus (APC)](<https://www.sota2.com/research/sota/role-play-faithfulness-on-prp-characters-with-persona-statements-at-scale-test.md?metric=Spartacus%20(APC)>): Higher is better; unit: SCALAR.

- [Hermione (APC)](<https://www.sota2.com/research/sota/role-play-faithfulness-on-prp-characters-with-persona-statements-at-scale-test.md?metric=Hermione%20(APC)>): Higher is better; unit: SCALAR.

- [Voldemort (APC)](<https://www.sota2.com/research/sota/role-play-faithfulness-on-prp-characters-with-persona-statements-at-scale-test.md?metric=Voldemort%20(APC)>): Higher is better; unit: SCALAR.

- [Cleopatra (APC)](<https://www.sota2.com/research/sota/role-play-faithfulness-on-prp-characters-with-persona-statements-at-scale-test.md?metric=Cleopatra%20(APC)>): Higher is better; unit: SCALAR.

- [Caesar (APC)](<https://www.sota2.com/research/sota/role-play-faithfulness-on-prp-characters-with-persona-statements-at-scale-test.md?metric=Caesar%20(APC)>): Higher is better; unit: SCALAR.

- [MLK (APC)](<https://www.sota2.com/research/sota/role-play-faithfulness-on-prp-characters-with-persona-statements-at-scale-test.md?metric=MLK%20(APC)>): Higher is better; unit: SCALAR.

- [Average (APC)](<https://www.sota2.com/research/sota/role-play-faithfulness-on-prp-characters-with-persona-statements-at-scale-test.md?metric=Average%20(APC)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Spartacus (APC) | Result date | Report paper |
| --- | --- | --- | --- |
| RAG | 70.1 | Not recorded | 68fa753e1e209131cea66b68 |
| RAG | 69.9 | Not recorded | 68fa753e1e209131cea66b68 |
| Gemma-7B | 69.4 | Not recorded | 68fa753e1e209131cea66b68 |
| ChatGPT | 69.1 | Not recorded | 68fa753e1e209131cea66b68 |
| GPT-4 | 69 | Not recorded | 68fa753e1e209131cea66b68 |
| Gemma-7B | 68.8 | Not recorded | 68fa753e1e209131cea66b68 |
| EU | 68.7 | Not recorded | 68fa753e1e209131cea66b68 |
| EU | 67.3 | Not recorded | 68fa753e1e209131cea66b68 |

### RAG

Spartacus (APC): 70.1. Result ID: 699397123a6c314849171652.

[Reporting paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Implementation](<https://github.com/KomeijiForce/Active_Passive_Constraint_Koishiday_2024>)

Recorded evaluation configuration: {"Configuration":"w/o CPO"}. This record does not assert independent reproduction.

### RAG

Spartacus (APC): 69.9. Result ID: 699397123a6c314849171655.

[Reporting paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Implementation](<https://github.com/KomeijiForce/Active_Passive_Constraint_Koishiday_2024>)

Recorded evaluation configuration: {"Configuration":"w/ CPO"}. This record does not assert independent reproduction.

### Gemma-7B

Spartacus (APC): 69.4. Result ID: 699397123a6c314849171653.

[Reporting paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Implementation](<https://github.com/KomeijiForce/Active_Passive_Constraint_Koishiday_2024>)

Recorded evaluation configuration: {"Configuration":"w/ CPO"}. This record does not assert independent reproduction.

### ChatGPT

Spartacus (APC): 69.1. Result ID: 699397123a6c31484917164e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

Recorded evaluation configuration: {"Configuration":"GPT baseline"}. This record does not assert independent reproduction.

### GPT-4

Spartacus (APC): 69. Result ID: 699397123a6c31484917164f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa727a1e209131cea56a1e.md>)

[Implementation](<https://github.com/openai/evals>)

Recorded evaluation configuration: {"Configuration":"GPT baseline"}. This record does not assert independent reproduction.

### Gemma-7B

Spartacus (APC): 68.8. Result ID: 699397123a6c314849171650.

[Reporting paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Implementation](<https://github.com/KomeijiForce/Active_Passive_Constraint_Koishiday_2024>)

Recorded evaluation configuration: {"Configuration":"w/o CPO"}. This record does not assert independent reproduction.

### EU

Spartacus (APC): 68.7. Result ID: 699397123a6c314849171654.

[Reporting paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Implementation](<https://github.com/KomeijiForce/Active_Passive_Constraint_Koishiday_2024>)

Recorded evaluation configuration: {"Configuration":"w/ CPO"}. This record does not assert independent reproduction.

### EU

Spartacus (APC): 67.3. Result ID: 699397123a6c314849171651.

[Reporting paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa753e1e209131cea66b68.md>)

[Implementation](<https://github.com/KomeijiForce/Active_Passive_Constraint_Koishiday_2024>)

Recorded evaluation configuration: {"Configuration":"w/o CPO"}. This record does not assert independent reproduction.

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