# Scientific Literature Embedding

BMRetriever is a 410-million-parameter large language model optimized for biomedical text retrieval tasks, with enhanced retrieval performance through fine-tuning

PubMedNCL is a pre-trained biomedical document representation language model based on PubMedBERT, fine-tuned through citation neighborhood contrastive learning.

Scientific literature embedding model trained without SciDocs leakage data

SciNCL is a pre-trained BERT language model for generating document-level embedding representations of research papers, trained using contrastive learning with citation graph neighborhood relationships.