Exploiting Language Characteristics for Legal Domain-Specific Language Model Pretraining

33261

Knowledge Triples

Source Entity	Relationship	Target Entity
Autoregressive Language Modelling Method	leverages	Legal Domain OtherEntity
Masked Language Modelling Method	leverages	Legal Domain OtherEntity
Pre-Training Method	improves	Legal Downstream Tasks Task
Pre-Training Method	includes	Autoregressive Language Modelling Method
Pre-Training Method	includes	Masked Language Modelling Method
Pre-Training Method	targets	Sentence Level Semantics Task
Pre-Training Method	targets	Token Level Feature Representation Task