LLM-generated text detection on MAGE DIPPER attack
77.44Human ScoreSearchLLM
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| SearchLLMBase Detector=DetectGPT2026.01 | 77.44 | 72.3 | 44.35 | — | — | — | |
| SearchLLMBase Detector=RADAR2026.01 | 76.86 | 79.39 | 56.03 | — | — | — | |
| SearchLLMBase Detector=Longformer2026.01 | 76.44 | 78.96 | 53.16 | — | — | — | |
| DetectGPTMethod Type=Base Detector2026.01 | 63.05 | 49.56 | 42.69 | — | — | — | |
| SearchLLMBase Detector=Binoculars2026.01 | 61.21 | 92.55 | 54.19 | — | — | — | |
| LongformerMethod Type=Base Detector2026.01 | 56.29 | 60.69 | 50.42 | — | — | — | |
| RADARMethod Type=Base Detector2026.01 | 53.06 | 64.84 | 53.09 | — | — | — | |
| BinocularsMethod Type=Base Detector2026.01 | 22.38 | 86.76 | 51.15 | — | — | — | |
| Binoculars2026.01 | — | — | — | 0.08 | 15.04 | 0.37 | |
| DetectGPT2026.01 | — | — | — | 1.77 | 0 | 0.44 | |
| Longformer2026.01 | — | — | — | 0.75 | 0.64 | 0.51 | |
| RADAR2026.01 | — | — | — | 0 | 3.76 | 0.32 | |
| SearchLLMComparison Context=Vs. Longformer2026.01 | — | — | — | 37.23 | 6.87 | 0.87 | |
| SearchLLMComparison Context=Vs. RADAR2026.01 | — | — | — | 37.23 | 10.43 | 0.87 | |
| SearchLLMComparison Context=Vs. DetectGPT2026.01 | — | — | — | 31.55 | 0 | 3.11 | |
| SearchLLMComparison Context=Vs. Binoculars2026.01 | — | — | — | 37.23 | 25.65 | 0.87 |