Benchmarks
TaskDatasetSOTA resultTrendResults
WeChat (test)
57.21Precision
29
WeChat (test)
1.898NLL
16
English-translated WeChat (test)
59.07Precision
14
WeChat English-translated (test)
64.75Precision
14
WeChat English-translated (test)
64.18Precision
14
WeChat real-domain (test)
60.53Fragment Order Consistency
10
WeChat
8.23P@10
8
WeChat (cold-start)
2.016NLL
5
WeChat (Full)
2.021NLL
5