Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

SelfInst

Benchmarks

Task NameDataset NameSOTA ResultTrend
Instruction FollowingSelfInst
Rouge-L21.7
73
Instruction FollowingSelfInst
R-L Score23.4
50
Instruction-tuningSelfInst
ROUGE-L21.31
21
Instruction Following EvaluationSelfInst Out-of-Distribution
GPT-4o Score51.6
17
GenerationSelfInst (test)
LLM-as-a-Judge Score60.16
2
Showing 5 of 5 rows