LooGLE
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
LooGLE
40Retrieval
21
Loogle SD
45.1Score
17
LooGLE
57CR Score
12
LooGLE Long Dependency QA
0.094BLEU-1
12
LooGLE ArXiv Paper Summarization
29.15BLEU-1
11
LOOGLE v2 (test)
78.7Overall Accuracy
10
LooGLE
57Reasoning Accuracy
10
LooGLE
27QA Accuracy
10
LooGLE (test)
84.2ShortQA Score
8
LooGLE
66.3EM
6
LooGLE
86.02Short QA Score
5
LooGLE CR 16k
19.78Score
5
LooGLE-MR 16k
15.1Score
5
LooGLE-SD 16k
45.1Score
5
LooGLE 900 (test)
82.8Success Rate
3