PRCA: Fitting Black-Box Large Language Models for Retrieval Question Answering via Pluggable Reward-Driven Contextual Adapter
Authors:
Preprint
Abstract
The Retrieval Question Answering (ReQA) task employs the retrieval-augmented framework, composed of a retriever and generator. The generator formulates the answer based on the documents retrieved by the retriever. Incorporating Large Language Models (LLMs) as generators is beneficial due to their advanced QA capabilities, but they are typically too large to be fine-tuned with budget constraints while some of them are only accessible via APIs. To tackle this issue and further improve ReQA performance, we propose a trainable Pluggable Reward-Driven Contextual Adapter (PRCA), keeping the generator as a black box. Positioned between the retriever and generator in a Pluggable manner, PRCA refines the retrieved information by operating in a token-autoregressive strategy via maximizing rewards of the reinforcement learning phase. Our experiments validate PRCA's effectiveness in enhancing ReQA performance on three datasets by up to 20% improvement to fit black-box LLMs into existing frameworks, demonstrating its considerable potential in the LLMs era.Keywords
All Related Versions
This publication has 15 references indexed in Scilit:
- REPLUG: Retrieval-Augmented Black-Box Language Models2024
- Query Rewriting in Retrieval-Augmented Large Language Models2023
- TopiOCQA: Open-domain Conversational Question Answering with Topic SwitchingTransactions of the Association for Computational Linguistics, 2022
- GLM: General Language Model Pretraining with Autoregressive Blank Infilling2022
- Retrieval Augmentation Reduces Hallucination in Conversation2021
- SimCSE: Simple Contrastive Learning of Sentence Embeddings2021
- Leveraging Passage Retrieval with Generative Models for Open Domain Question Answering2021
- Dense Passage Retrieval for Open-Domain Question Answering2020
- BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension2020
- The Probabilistic Relevance Framework: BM25 and BeyondFoundations and Trends® in Information Retrieval, 2009