List your company
ResearchFormerly WizWand

Multiclass text classification on Twitter-Fin

91.87AccuracyLS-LLaMA-2-7B
20.661239.148157.63576.1219Oct 2, 2023
Updated 8mo ago

Evaluation results

MethodLinks
LS-LLaMA-2-7BModel Size=7B2023.1091.87
LS-unLLaMA-2-7BModel Size=7B2023.1091.54
LS-LLaMA-2-13BModel Size=13B2023.1091.2
RoBERTa-LargeModel Architecture=Discriminant baseline2023.1090.95
RoBERTa-BaseModel Architecture=Discriminant baseline2023.1090.32
BERT-LargeModel Architecture=Discriminant baseline2023.1088.74
BERT-BaseModel Architecture=Discriminant baseline2023.1088.19
LS-unLLaMA-2-13BModel Size=13B2023.1087.94
LLaMA-2-7BEvaluation Protocol=instruction-tuning, Model Size=7B2023.1068.72
LLaMA-2-13BEvaluation Protocol=zero-shot, Model Size=13B2023.1038.74
LLaMA-2-7BEvaluation Protocol=zero-shot, Model Size=7B2023.1023.4