List your company
ResearchFormerly WizWand

Machine-generated text detection on SQuAD

99AUROCOurs
34.5251.266884.74Jan 4, 2025
Updated 4mo ago

Evaluation results

MethodLinks
OursGenerator=LLAMA 3 8B, Evaluation Protocol=Zero-shot2025.0199——
OursGenerator=GPT-NEOX Erebus, Evaluation Protocol=Zero-shot2025.0199——
OursGenerator=LLAMA 3 8B, Evaluation Protocol=Supervised2025.0199——
OursGenerator=GPT-NEOX Erebus, Evaluation Protocol=Supervised2025.0199——
OursGenerator=QWEN 32B, Evaluation Protocol=Zero-shot2025.0197——
Roberta (base)Generator=LLAMA 3 8B, Evaluation Protocol=Supervised2025.0197——
OursGenerator=QWEN 32B, Evaluation Protocol=Supervised2025.0197——
Roberta (large)Generator=LLAMA 3 8B, Evaluation Protocol=Supervised2025.0195——
LogRankGenerator=LLAMA 3 8B, Evaluation Protocol=Zero-shot2025.0193——
Roberta (large)Generator=GPT-NEOX Erebus, Evaluation Protocol=Supervised2025.0193——
Roberta (base)Generator=GPT-NEOX Erebus, Evaluation Protocol=Supervised2025.0192——
log p(x)Generator=LLAMA 3 8B, Evaluation Protocol=Zero-shot2025.0191——
LogRankGenerator=GPT-NEOX Erebus, Evaluation Protocol=Zero-shot2025.0181——
DetectGPTGenerator=GPT-NEOX Erebus, Evaluation Protocol=Zero-shot2025.0178——
BinocularsGenerator=GPT-NEOX Erebus, Evaluation Protocol=Zero-shot2025.0178——
log p(x)Generator=GPT-NEOX Erebus, Evaluation Protocol=Zero-shot2025.0175——
RankGenerator=LLAMA 3 8B, Evaluation Protocol=Zero-shot2025.0171——
RankGenerator=GPT-NEOX Erebus, Evaluation Protocol=Zero-shot2025.0171——
EntropyGenerator=QWEN 32B, Evaluation Protocol=Zero-shot2025.0170——
Roberta (base)Generator=QWEN 32B, Evaluation Protocol=Supervised2025.0169——
Roberta (large)Generator=QWEN 32B, Evaluation Protocol=Supervised2025.0168——
EntropyGenerator=GPT-NEOX Erebus, Evaluation Protocol=Zero-shot2025.0166——
LogRankGenerator=QWEN 32B, Evaluation Protocol=Zero-shot2025.0162——
DetectGPTGenerator=QWEN 32B, Evaluation Protocol=Zero-shot2025.0162——
BinocularsGenerator=QWEN 32B, Evaluation Protocol=Zero-shot2025.0162——
log p(x)Generator=QWEN 32B, Evaluation Protocol=Zero-shot2025.0160——
RankGenerator=QWEN 32B, Evaluation Protocol=Zero-shot2025.0155——
DetectGPTGenerator=LLAMA 3 8B, Evaluation Protocol=Zero-shot2025.0155——
BinocularsGenerator=LLAMA 3 8B, Evaluation Protocol=Zero-shot2025.0155——
EntropyGenerator=LLAMA 3 8B, Evaluation Protocol=Zero-shot2025.0137——
BinocularsMethod Category=Metric-based methods, Evaluation Scenario=IND-MGT2025.11—79.6579.29
DEERMethod Category=Domain generalization methods, Evaluation Scenario=IND-MGT2025.11—93.1593.2
EAGLEMethod Category=Model-based methods, Evaluation Scenario=IND-MGT2025.11—92.0892.39
EntropyMethod Category=Metric-based methods, Evaluation Scenario=IND-MGT2025.11—67.5163.79
Fast-DetectGPTMethod Category=Metric-based methods, Evaluation Scenario=IND-MGT2025.11—75.6576.27
GhostbusterMethod Category=Model-based methods, Evaluation Scenario=IND-MGT2025.11—89.9989.89
ImBDMethod Category=Model-based methods, Evaluation Scenario=IND-MGT2025.11—74.8779.65
Log-LikelihoodMethod Category=Metric-based methods, Evaluation Scenario=IND-MGT2025.11—77.9878.63
LRRMethod Category=Metric-based methods, Evaluation Scenario=IND-MGT2025.11—7572.49
MGT-PrismMethod Category=Domain generalization methods, Evaluation Scenario=IND-MGT2025.11—91.8192.57
MoSEsMethod Category=Model-based methods, Evaluation Scenario=IND-MGT2025.11—86.386.16
MSCLMethod Category=Domain generalization methods, Evaluation Scenario=IND-MGT2025.11—91.691.86
PeCoLAMethod Category=Model-based methods, Evaluation Scenario=IND-MGT2025.11—91.6791.65
RankMethod Category=Metric-based methods, Evaluation Scenario=IND-MGT2025.11—68.864.77
RobertaMethod Category=Model-based methods, Evaluation Scenario=IND-MGT2025.11—92.492.74
TACITMethod Category=Domain generalization methods, Evaluation Scenario=IND-MGT2025.11—86.487.09