MetaQA
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
MetaQA 1-hop LLaMA2-7B responses (val)
83.41AUC
7
MetaQA 1hop (LLaMA2-7B)
83.41AUC
7
MetaQA 1-hop wikimovies
97.5Hits@1 (KB)
6
MetaQA 3-hop (test)
55.31Recall
4
MetaQA 2-hop (test)
91.9Recall
4
MetaQA 1-hop (test)
98.86Recall
4
MetaQA 3hop
1Steps
4
MetaQA 2hop
1Steps
4
MetaQA
95.5HS
4
MetaQA 3-hop
89Comprehensiveness
3
MetaQA 2-hop
86Comprehensiveness
3
MetaQA 1-hop
81Comprehensiveness
3
MetaQA
8,321Detected Samples
3
MetaQA (test)
96.6HS
2
MetaQA
100CDPA
1