Machine-Generated Text Detection on Reuters ChatGPT split benchmark leaderboard · SOTA2 Research

Machine-Generated Text Detection on Reuters ChatGPT split

98.09TPR@FPR=1%DNA-GPT-M

Evaluation Results

Method	Links
DNA-GPT-M2026.02		98.09	99.84
DNA-GPT-M2026.02		96.76	99.78
DNA-GPT2026.02		91.47	99.74
Log-Rank-M2026.02		90.62	98.69
Log-Rank-M2026.02		90.49	98.71
Likelihood-M2026.02		88.98	98.68
Likelihood-M2026.02		88.58	98.45
Log-Rank2026.02		86.22	97.77
Likelihood2026.02		81.07	97.55
Likelihood2026.02		81.07	97.55
DNA-GPT2026.02		69.78	99.17
DetectGPT-M2026.02		47.91	92.37
DetectGPT-M2026.02		37.2	73.22
FastGPT-M2026.02		21.78	90.23
Entropy-M2026.02		17.47	80.07
Log-Rank2026.02		17.24	21.28
FastGPT-M2026.02		17.07	89.97
Entropy-M2026.02		7.29	56.16
Entropy2026.02		2.49	64.65
FastGPT2026.02		2.04	60.44
Entropy2026.02		0.04	35.35
FastGPT2026.02		0.04	39.56
DetectGPT2026.02		0	57.18
DetectGPT2026.02		0	57.18