RAFT

Adapting Language Model to Domain Specific RAG

Pretraining Large Language Models (LLMs) on large corpora of textual data is now a standard paradigm. When using these LLMs for many downstream applications, it is common to additionally bake in new knowledge (e.g., timecritical news, or private domain knowledge) into the pretrained model either through RAG-basedprompting, or finetuning. However, the optimal methodology for the model to gain such new knowledge remains an open question. In this paper, we present Retrieval Augmented Fine Tuning (RAFT), a training recipe that improves the model’s ability to answer questions in an “openbook” in-domain setting. In RAFT, given a question, and a set of retrieved documents, we train the model to ignore those documents that don’t help in answering the question, which we call, distractor documents. RAFT accomplishes this by citing verbatim the right sequence from the relevant document that would help answer the question. This coupled with RAFT’s chain-of-thoughtstyle response helps improve the model’s ability to reason. In domain specific RAG, RAFT consistently improves the model’s performance across PubMed, HotpotQA, and Gorilla datasets, presenting a post-training recipe to improve pre-trained LLMs to in-domain RAG.

---