General Domain
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
General Domain
31.27Retrieval Score
30
General Domain 31 tasks
57.97CLS Score
30
General Domain AlpacaEval Arena-Hard LLaMA3-8B (10% selection)
12.09AlpacaEval Score
18
General-domain (test)
40.77BLEU
6
General Domain Average
42.35Average EM
5
General Domain (holdout test)
0.79L_inf
4