L-Eval
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
L-Eval
58.28Coursera
26
L-Eval (test)
58.28Coursera
26
L-Eval Sum
22.66QMS
13
L-Eval QA
80.73NQ
13
L-Eval
68.8Close Score
13
L-Eval 32K
215P95 Latency (ms)
12
L-Eval closed-ended tasks
41.86Coursera Score
12
L-Eval Closed-ended
48.51Coursera Score
6
L-Eval
70.6Coursera Accuracy
6
L-Eval (test)
64.4Coursera QA Accuracy
5
L-Eval
30.2Coursera QA
4