RAGTruth
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
RAGTruth (test)
0.91AUROC
99
RAGTruth
0.854AUROC
79
RAGTruth CNN/DM (subsample)
0.69AUROC
45
RAGTruth MS MARCO (subsample)
0.77AUROC
45
RAGTruth RT-QA 1.0 (test)
0.789F1 Score
33
RAGTruth RT-Summ 1.0 (test)
0.697F1 Score
30
RAGTruth RT-D2T 1.0 (test)
0.738F1 Score
30
RAGTruth Llama2-13B (test)
83.33Acc
21
RAGTruth Llama2-7B (test)
75.76Accuracy
21
RAGTruth LLaMA3-8B
78.6Recall
19
RAGTruth LLaMA2-13B
80.68Recall
19
RAGTruth LLaMA2-7B
0.833Recall
19
RAGTruth
59.4AP (Token-level)
18
RAGTruth
75.96AP
18
RAGTruth summarization (test)
52ROUGE-1
18
RAGTruth
45.89F1 Score
17
RAGTruth (test)
74.5AUC
15
RAGTruth Span-level, leakage-clean protocol
0.702AUC
15
RAGTruth summarization task
77Precision
14
RAGTruth QA
91.89AUROC
13
RAGTruth
98.4Faithfulness
12
RagTruth-Avg (test)
76.63F1 Score
12
RAGTruth
33.14F1 Score
11
RagTruth
0.57Kendall's Tau
11
RAGTruth
90.3Accuracy
10