# Role-playing Dialogue Evaluation on FURINA-Bench Chinese

Source: [Role-playing Dialogue Evaluation on FURINA-Bench Chinese](<https://www.sota2.com/research/sota/role-playing-dialogue-evaluation-on-furina-bench-chinese>)

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

Data updated: 2026-04-07T09:32:38.586Z

Evaluation of model performance on the FURINA-Bench Chinese dataset, focusing on role-playing dialogue capabilities across dimensions such as Context Reliance, Factual Recall, Reflective Reasoning, Conversational Ability, and Preference Alignment.

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 Dialogue Evaluation |
| Dataset (including any stated split) | FURINA-Bench Chinese |
| Metric | Context Reliance |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 12 of 12 |
| 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

- [Context Reliance](<https://www.sota2.com/research/sota/role-playing-dialogue-evaluation-on-furina-bench-chinese.md?metric=Context%20Reliance>): Higher is better; unit: PERCENT.

- [Factual Recall](<https://www.sota2.com/research/sota/role-playing-dialogue-evaluation-on-furina-bench-chinese.md?metric=Factual%20Recall>): Higher is better; unit: PERCENT.

- [Reflective Reasoning](<https://www.sota2.com/research/sota/role-playing-dialogue-evaluation-on-furina-bench-chinese.md?metric=Reflective%20Reasoning>): Higher is better; unit: PERCENT.

- [Conversational Ability](<https://www.sota2.com/research/sota/role-playing-dialogue-evaluation-on-furina-bench-chinese.md?metric=Conversational%20Ability>): Higher is better; unit: PERCENT.

- [Preference Alignment](<https://www.sota2.com/research/sota/role-playing-dialogue-evaluation-on-furina-bench-chinese.md?metric=Preference%20Alignment>): Higher is better; unit: PERCENT.

- [Weighted Average Score](<https://www.sota2.com/research/sota/role-playing-dialogue-evaluation-on-furina-bench-chinese.md?metric=Weighted%20Average%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Context Reliance | Result date | Report paper |
| --- | --- | --- | --- |
| Qwen3-32B | 71.39 | Not recorded | 68fa7c101e209131cea8c38e |
| Deepseek-R1 | 67.91 | Not recorded | 68fa7c101e209131cea8c38e |
| Qwen3-235B | 67.07 | Not recorded | 68fa7c101e209131cea8c38e |
| Qwen3-32B | 64.14 | Not recorded | 68fa7c101e209131cea8c38e |
| Qwen3-235B | 59.38 | Not recorded | 68fa7c101e209131cea8c38e |
| Qwen3-8B | 53.23 | Not recorded | 68fa7c101e209131cea8c38e |
| o3 | 47.46 | Not recorded | 68fa7c101e209131cea8c38e |
| Qwen3-8B | 45.07 | Not recorded | 68fa7c101e209131cea8c38e |
| Claude-4-Sonnet | 40.17 | Not recorded | 68fa7c101e209131cea8c38e |
| Claude-4-Sonnet | 37.77 | Not recorded | 68fa7c101e209131cea8c38e |
| Deepseek-V3 | 36.31 | Not recorded | 68fa7c101e209131cea8c38e |
| GPT-4o | 28.49 | Not recorded | 68fa7c101e209131cea8c38e |

### Qwen3-32B

Context Reliance: 71.39. Result ID: 69d4cf3547f34ce924bceeb6.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"thinking":"true"}. This record does not assert independent reproduction.

### Deepseek-R1

Context Reliance: 67.91. Result ID: 69d4cf3547f34ce924bceebb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76a91e209131cea6e3eb.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Qwen3-235B

Context Reliance: 67.07. Result ID: 69d4cf3547f34ce924bceeb8.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"thinking":"true"}. This record does not assert independent reproduction.

### Qwen3-32B

Context Reliance: 64.14. Result ID: 69d4cf3547f34ce924bceeb5.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"thinking":"false"}. This record does not assert independent reproduction.

### Qwen3-235B

Context Reliance: 59.38. Result ID: 69d4cf3547f34ce924bceeb7.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"thinking":"false"}. This record does not assert independent reproduction.

### Qwen3-8B

Context Reliance: 53.23. Result ID: 69d4cf3547f34ce924bceeb4.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"thinking":"true"}. This record does not assert independent reproduction.

### o3

Context Reliance: 47.46. Result ID: 69d4cf3547f34ce924bceebe.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Qwen3-8B

Context Reliance: 45.07. Result ID: 69d4cf3547f34ce924bceeb3.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa788e1e209131cea79116.md>)

Recorded evaluation configuration: {"thinking":"false"}. This record does not assert independent reproduction.

### Claude-4-Sonnet

Context Reliance: 40.17. Result ID: 69d4cf3547f34ce924bceebd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

Recorded evaluation configuration: {"thinking":"true"}. This record does not assert independent reproduction.

### Claude-4-Sonnet

Context Reliance: 37.77. Result ID: 69d4cf3547f34ce924bceebc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

Recorded evaluation configuration: {"thinking":"false"}. This record does not assert independent reproduction.

### Deepseek-V3

Context Reliance: 36.31. Result ID: 69d4cf3547f34ce924bceeba.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa77001e209131cea7066a.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### GPT-4o

Context Reliance: 28.49. Result ID: 69d4cf3547f34ce924bceeb9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa7c101e209131cea8c38e.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

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