LLM-generated text detection on RAID Wikipedia-related (all)
89.12GPT-4 ScoreSearchLLM
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SearchLLMBase Detector=Longformer2026.01 | 89.12 | 92.97 | 61.02 | 78.57 | 84.56 | 81.24 | |
| Binoculars2026.01 | 77.36 | 5.61 | 92.33 | 5.61 | 50.46 | 46.27 | |
| SearchLLMBase Detector=Binoculars2026.01 | 72.07 | 61.82 | 87.54 | 61.82 | 67.32 | 70.11 | |
| SearchLLMBase Detector=RADAR2026.01 | 71.66 | 82.68 | 48.63 | 82.35 | 77.58 | 72.58 | |
| RADAR2026.01 | 61.9 | 74.93 | 42.41 | 76.28 | 70.03 | 65.11 | |
| Longformer2026.01 | 48.24 | 48.2 | 25.05 | 39.24 | 40.08 | 40.16 | |
| SearchLLMBase Detector=DetectGPT2026.01 | 6.39 | 8.96 | 2.88 | 5.02 | 8.22 | 6.29 | |
| DetectGPT2026.01 | 1.74 | 3.4 | 0.97 | 1.86 | 2.92 | 2.18 |