Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Peer-Review

Benchmarks

Task NameDataset NameSOTA ResultTrend
AI-generated text detectionPeer-Review Claude generated
TPR@FPR=0.1%64.02
14
AI-generated text detectionPeer-Review Gemini generated
TPR @ FPR=0.1%43.96
14
AI-generated text detectionPeer-Review GPT-4o generated
TPR @ 0.1% FPR33.04
14
Adversarial attack on AI-text detectorsPeer-review (evaluation set)
RoBERTa ASR63
12
AI Text DetectionPeer-Review
TPR@0.1%67.36
6
Machine-generated text detectionPeer-Review Claude
AUROC99.4
5
Machine-generated text detectionPeer-Review Gemini
AUROC97.9
5
Machine-generated text detectionPeer-Review GPT-4o
AUROC0.973
5
Showing 8 of 8 rows