LLM-generated content detection on HumanEval
4.87Std Dev of F1 ScoreDSIPA (SDP)
Evaluation results
| Method | Links | |
|---|---|---|
| DSIPA (SDP)variant=SDP2026.04 | 4.87 | |
| RoBERTa-basemodel size=base2026.04 | 4.92 | |
| R-Detect2026.04 | 4.93 | |
| Binoculars2026.04 | 5.09 | |
| DSIPA (SDC)variant=SDC2026.04 | 5.16 | |
| Fast-DetectGPT2026.04 | 5.21 | |
| RAIDAR2026.04 | 5.42 | |
| GPT-Zero2026.04 | 5.53 | |
| RoBERTa-largemodel size=large2026.04 | 5.88 | |
| DetectGPT2026.04 | 5.97 | |
| Ghostbuster2026.04 | 6.05 | |
| LogRank2026.04 | 6.21 |