1 paper · 2 filters
Yucheng Cai, Si Chen, Yuxuan Wu +3
Recently, increasing research interests have focused on retrieval augmented generation (RAG) to mitigate hallucination for large language models (LLMs). Following this trend, we la…