# Machine-generated text detection on GPT3.5 and HC3 Cross-generator 64-shot

Source: [Machine-generated text detection on GPT3.5 and HC3 Cross-generator 64-shot](<https://www.sota2.com/research/sota/machine-generated-text-detection-on-gpt3-5-and-hc3-cross-generator-64-shot>)

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

Data updated: 2026-02-26T05:45:47.518Z

Evaluation of machine-generated text detection models using the cross-generator setup between GPT-3.5 and HC3 datasets, specifically using a 64-shot prompting configuration.

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 | Machine-generated text detection |
| Dataset (including any stated split) | GPT3.5 and HC3 Cross-generator 64-shot |
| Metric | GPT3.5 -\> HC3 Score |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [GPT3.5 -\> HC3 Score](<https://www.sota2.com/research/sota/machine-generated-text-detection-on-gpt3-5-and-hc3-cross-generator-64-shot.md?metric=GPT3.5%20-%3E%20HC3%20Score>): Higher is better; unit: PERCENT.

- [HC3 -\> GPT3.5 Score](<https://www.sota2.com/research/sota/machine-generated-text-detection-on-gpt3-5-and-hc3-cross-generator-64-shot.md?metric=HC3%20-%3E%20GPT3.5%20Score>): Higher is better; unit: PERCENT.

- [Average Score](<https://www.sota2.com/research/sota/machine-generated-text-detection-on-gpt3-5-and-hc3-cross-generator-64-shot.md?metric=Average%20Score>): Higher is better; unit: PERCENT.

## Recorded results

| Method | GPT3.5 -\> HC3 Score | Result date | Report paper |
| --- | --- | --- | --- |
| PECOLA | 78.79 | Not recorded | 68fa73fd1e209131cea5f0f0 |
| DetectGPT | 77.11 | Not recorded | 68fa73fd1e209131cea5f0f0 |
| RoBERTa | 64.6 | Not recorded | 68fa73fd1e209131cea5f0f0 |

### PECOLA

GPT3.5 -\> HC3 Score: 78.79. Result ID: 699313473a6c31484913848c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73fd1e209131cea5f0f0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73fd1e209131cea5f0f0.md>)

[Implementation](<https://github.com/lsc-1/Pecola>)

Recorded evaluation configuration: {"Setting":"64-shot"}. This record does not assert independent reproduction.

### DetectGPT

GPT3.5 -\> HC3 Score: 77.11. Result ID: 699313473a6c31484913848b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73fd1e209131cea5f0f0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa701a1e209131cea49064.md>)

Recorded evaluation configuration: {"Setting":"64-shot"}. This record does not assert independent reproduction.

### RoBERTa

GPT3.5 -\> HC3 Score: 64.6. Result ID: 699313473a6c31484913848a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa73fd1e209131cea5f0f0.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66131e209131cea0e18b.md>)

[Implementation](<https://github.com/pytorch/fairseq>)

Recorded evaluation configuration: {"Setting":"64-shot"}. This record does not assert independent reproduction.

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