LLM-generated text detection on DetectRL Out-of-Domain Multi-Topic 1.0 (test)
91.1Average Detection ScoreBinoculars + MoP
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Binoculars + MoPSource Proxy=w/o src2026.04 | 91.1 | 88 | 93.3 | 90.1 | 93.1 | |
| Binoculars + kNNProxySource Proxy=arXiv2026.04 | 90.9 | — | 91.8 | 87.8 | 90.7 | |
| Binoculars + kNNProxySource Proxy=XSum2026.04 | 90.9 | 90.2 | — | 90.7 | 93.2 | |
| Binoculars + kNNProxySource Proxy=Writing2026.04 | 90.9 | 88.2 | 93.9 | — | 93.1 | |
| Binoculars + kNNProxySource Proxy=Review2026.04 | 90.9 | 87.7 | 93.5 | 90.1 | — | |
| Fast-DetectGPT + MoPSource Proxy=w/o src2026.04 | 79.8 | 73.8 | 74 | 84.4 | 86.9 | |
| Fast-DetectGPT + kNNProxySource Proxy=arXiv2026.04 | 79.5 | — | 73.7 | 81.3 | 86.5 | |
| Fast-DetectGPT + kNNProxySource Proxy=XSum2026.04 | 79.5 | 75.1 | — | 84.9 | 87.6 | |
| Fast-DetectGPT + kNNProxySource Proxy=Writing2026.04 | 79.5 | 73.1 | 74.1 | — | 86.8 | |
| Fast-DetectGPT + kNNProxySource Proxy=Review2026.04 | 79.5 | 72.8 | 73.3 | 84.5 | — | |
| BinocularsSource Proxy=arXiv2026.04 | 77.4 | — | 72 | 79.5 | 83.6 | |
| BinocularsSource Proxy=XSum2026.04 | 77.4 | 76.3 | — | 79.7 | 83.2 | |
| BinocularsSource Proxy=Writing2026.04 | 77.4 | 73.6 | 69.7 | — | 81.8 | |
| BinocularsSource Proxy=Review2026.04 | 77.4 | 77.7 | 72 | 79.4 | — | |
| Fast-DetectGPTSource Proxy=arXiv2026.04 | 48.4 | — | 23.7 | 59.7 | 60.2 | |
| Fast-DetectGPTSource Proxy=XSum2026.04 | 48.4 | 28.4 | — | 63 | 63.1 | |
| Fast-DetectGPTSource Proxy=Writing2026.04 | 48.4 | 34.8 | 33.6 | — | 68.3 | |
| Fast-DetectGPTSource Proxy=Review2026.04 | 48.4 | 40.7 | 37.7 | 68.3 | — |