# Content Generation on CoGenesis human assessment and automated evaluation

Source: [Content Generation on CoGenesis human assessment and automated evaluation](<https://www.sota2.com/research/sota/content-generation-on-cogenesis-human-assessment-and-automated-evaluation>)

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

Data updated: 2026-02-26T06:06:26.734Z

Evaluation of content generation models using a combination of human preference metrics (Win/Tie/Lose) and automated metrics (BLEU, ROUGE-L) on the CoGenesis dataset.

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 | Content Generation |
| Dataset (including any stated split) | CoGenesis human assessment and automated evaluation |
| Metric | Tie Rate |
| 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

- [Tie Rate](<https://www.sota2.com/research/sota/content-generation-on-cogenesis-human-assessment-and-automated-evaluation.md?metric=Tie%20Rate>): Higher is better; unit: PERCENT.

- [Win Rate](<https://www.sota2.com/research/sota/content-generation-on-cogenesis-human-assessment-and-automated-evaluation.md?metric=Win%20Rate>): Higher is better; unit: PERCENT.

- [Loss Rate](<https://www.sota2.com/research/sota/content-generation-on-cogenesis-human-assessment-and-automated-evaluation.md?metric=Loss%20Rate>): Lower is better; unit: PERCENT.

- [BLEU](<https://www.sota2.com/research/sota/content-generation-on-cogenesis-human-assessment-and-automated-evaluation.md?metric=BLEU>): Higher is better; unit: SCALAR.

- [ROUGE-L](<https://www.sota2.com/research/sota/content-generation-on-cogenesis-human-assessment-and-automated-evaluation.md?metric=ROUGE-L>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Tie Rate | Result date | Report paper |
| --- | --- | --- | --- |
| SLM finetuned | 50 | Not recorded | 68fa738b1e209131cea5d1d0 |
| Logits-based CoGen | 15 | Not recorded | 68fa738b1e209131cea5d1d0 |
| Sketch-based CoGen | 13 | Not recorded | 68fa738b1e209131cea5d1d0 |
| LLM w/ context | 12 | Not recorded | 68fa738b1e209131cea5d1d0 |
| LLM w/o context | 10 | Not recorded | 68fa738b1e209131cea5d1d0 |

### SLM finetuned

Tie Rate: 50. Result ID: 6992ef833a6c314849127814.

[Reporting paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1d0.md>)

Recorded evaluation configuration: {"Model and Setting":"SLM finetuned","Base Model":"Qwen-1.8B"}. This record does not assert independent reproduction.

### Logits-based CoGen

Tie Rate: 15. Result ID: 6992ef833a6c314849127818.

[Reporting paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1d0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1d0.md>)

Recorded evaluation configuration: {"Model and Setting":"Logits-based CoGen","LLM":"Qwen-72B","SLM":"Qwen-1.8B"}. This record does not assert independent reproduction.

### Sketch-based CoGen

Tie Rate: 13. Result ID: 6992ef833a6c314849127817.

[Reporting paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1d0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1d0.md>)

Recorded evaluation configuration: {"Model and Setting":"Sketch-based CoGen","LLM":"Qwen-72B","SLM":"Qwen-1.8B"}. This record does not assert independent reproduction.

### LLM w/ context

Tie Rate: 12. Result ID: 6992ef833a6c314849127815.

[Reporting paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1d0.md>)

Recorded evaluation configuration: {"Model and Setting":"LLM w/ context","Base Model":"Qwen-72B"}. This record does not assert independent reproduction.

### LLM w/o context

Tie Rate: 10. Result ID: 6992ef833a6c314849127816.

[Reporting paper](<https://www.sota2.com/research/paper/68fa738b1e209131cea5d1d0.md>)

Recorded evaluation configuration: {"Model and Setting":"LLM w/o context","Base Model":"Qwen-72B"}. This record does not assert independent reproduction.

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