# Role-Playing on RPGBench In-distribution

Source: [Role-Playing on RPGBench In-distribution](<https://www.sota2.com/research/sota/role-playing-on-rpgbench-in-distribution>)

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

Data updated: 2026-02-26T05:59:13.770Z

Evaluation of role-playing capabilities using the RPGBench In-distribution dataset, measured by the R-EMI metric.

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-Playing |
| Dataset (including any stated split) | RPGBench In-distribution |
| Metric | R-EMI |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 18 of 18 |
| 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

- [R-EMI](<https://www.sota2.com/research/sota/role-playing-on-rpgbench-in-distribution.md?metric=R-EMI>): Higher is better; unit: SCALAR.

## Recorded results

| Method | R-EMI | Result date | Report paper |
| --- | --- | --- | --- |
| RL | -0.034 | Not recorded | 6948d057423ea42265729d41 |
| CoRL | -0.04 | Not recorded | 6948d057423ea42265729d41 |
| CoRL | -0.049 | Not recorded | 6948d057423ea42265729d41 |
| RL | -0.05 | Not recorded | 6948d057423ea42265729d41 |
| SFT | -0.052 | Not recorded | 6948d057423ea42265729d41 |
| ICL | -0.065 | Not recorded | 6948d057423ea42265729d41 |
| SFT | -0.067 | Not recorded | 6948d057423ea42265729d41 |
| DA | -0.079 | Not recorded | 6948d057423ea42265729d41 |
| DA | -0.09 | Not recorded | 6948d057423ea42265729d41 |
| ICL | -0.096 | Not recorded | 6948d057423ea42265729d41 |
| ThinkingSFT | -0.107 | Not recorded | 6948d057423ea42265729d41 |
| ThinkingSFT | -0.121 | Not recorded | 6948d057423ea42265729d41 |
| ThinkingRL | -0.196 | Not recorded | 6948d057423ea42265729d41 |
| ThinkingRL | -0.272 | Not recorded | 6948d057423ea42265729d41 |
| DPO | -0.287 | Not recorded | 6948d057423ea42265729d41 |
| RFT | -0.303 | Not recorded | 6948d057423ea42265729d41 |
| DPO | -0.586 | Not recorded | 6948d057423ea42265729d41 |
| RFT | -0.835 | Not recorded | 6948d057423ea42265729d41 |

### RL

R-EMI: -0.034. Result ID: 6999501b5698665ec4ba9a15.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73101e209131cea5a395.md>)

[Implementation](<https://github.com/deepseek-ai/DeepSeek-Math>)

Recorded evaluation configuration: {"Backbone":"Meta-Llama3-8B"}. This record does not assert independent reproduction.

### CoRL

R-EMI: -0.04. Result ID: 6999501b5698665ec4ba9a0f.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### CoRL

R-EMI: -0.049. Result ID: 6999501b5698665ec4ba9a18.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

Recorded evaluation configuration: {"Backbone":"Meta-Llama3-8B"}. This record does not assert independent reproduction.

### RL

R-EMI: -0.05. Result ID: 6999501b5698665ec4ba9a0c.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73101e209131cea5a395.md>)

[Implementation](<https://github.com/deepseek-ai/DeepSeek-Math>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### SFT

R-EMI: -0.052. Result ID: 6999501b5698665ec4ba9a11.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

Recorded evaluation configuration: {"Backbone":"Meta-Llama3-8B"}. This record does not assert independent reproduction.

### ICL

R-EMI: -0.065. Result ID: 6999501b5698665ec4ba9a07.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73a71e209131cea5db60.md>)

[Implementation](<https://github.com/InteractiveNLP-Team/RoleLLM-public>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### SFT

R-EMI: -0.067. Result ID: 6999501b5698665ec4ba9a08.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### DA

R-EMI: -0.079. Result ID: 6999501b5698665ec4ba9a12.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76e01e209131cea6f891.md>)

Recorded evaluation configuration: {"Backbone":"Meta-Llama3-8B"}. This record does not assert independent reproduction.

### DA

R-EMI: -0.09. Result ID: 6999501b5698665ec4ba9a09.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76e01e209131cea6f891.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### ICL

R-EMI: -0.096. Result ID: 6999501b5698665ec4ba9a10.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73a71e209131cea5db60.md>)

[Implementation](<https://github.com/InteractiveNLP-Team/RoleLLM-public>)

Recorded evaluation configuration: {"Backbone":"Meta-Llama3-8B"}. This record does not assert independent reproduction.

### ThinkingSFT

R-EMI: -0.107. Result ID: 6999501b5698665ec4ba9a16.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79021e209131cea7ba49.md>)

Recorded evaluation configuration: {"Backbone":"Meta-Llama3-8B"}. This record does not assert independent reproduction.

### ThinkingSFT

R-EMI: -0.121. Result ID: 6999501b5698665ec4ba9a0d.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79021e209131cea7ba49.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### ThinkingRL

R-EMI: -0.196. Result ID: 6999501b5698665ec4ba9a0e.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79021e209131cea7ba49.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### ThinkingRL

R-EMI: -0.272. Result ID: 6999501b5698665ec4ba9a17.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79021e209131cea7ba49.md>)

Recorded evaluation configuration: {"Backbone":"Meta-Llama3-8B"}. This record does not assert independent reproduction.

### DPO

R-EMI: -0.287. Result ID: 6999501b5698665ec4ba9a0a.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73f21e209131cea5f078.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

### RFT

R-EMI: -0.303. Result ID: 6999501b5698665ec4ba9a14.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

Recorded evaluation configuration: {"Backbone":"Meta-Llama3-8B"}. This record does not assert independent reproduction.

### DPO

R-EMI: -0.586. Result ID: 6999501b5698665ec4ba9a13.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73f21e209131cea5f078.md>)

Recorded evaluation configuration: {"Backbone":"Meta-Llama3-8B"}. This record does not assert independent reproduction.

### RFT

R-EMI: -0.835. Result ID: 6999501b5698665ec4ba9a0b.

[Reporting paper](<https://www.sota2.com/research/paper/6948d057423ea42265729d41.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-4B"}. This record does not assert independent reproduction.

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