Passage-specific Prompt Tuning for Passage Reranking in Question Answering with Large Language Models
Wu, Xuyang, Peng, Zhiyuan, Sai, Krishna Sravanthi Rajanala, Wu, Hsin-Tai, Fang, Yi
–arXiv.org Artificial Intelligence
Effective passage retrieval and reranking methods have been widely utilized to identify suitable candidates in open-domain question answering tasks, recent studies have resorted to LLMs for reranking the retrieved passages by the log-likelihood of the question conditioned on each passage. Although these methods have demonstrated promising results, the performance is notably sensitive to the human-written prompt (or hard prompt), and fine-tuning LLMs can be computationally intensive and time-consuming. Furthermore, this approach limits the leverage of question-passage relevance pairs and passage-specific knowledge to enhance the ranking capabilities of LLMs. In this paper, we propose passage-specific prompt tuning for reranking in open-domain question answering (PSPT): a parameter-efficient method that fine-tunes learnable passage-specific soft prompts, incorporating passage-specific knowledge from a limited set of question-passage relevance pairs. The method involves ranking retrieved passages based on the log-likelihood of the model generating the question conditioned on each passage and the learned soft prompt. We conducted extensive experiments utilizing the Llama-2-chat-7B model across three publicly available open-domain question answering datasets and the results demonstrate the effectiveness of the proposed approach.
arXiv.org Artificial Intelligence
Jun-20-2024
- Country:
- North America > United States
- District of Columbia > Washington (0.05)
- Maryland (0.04)
- Louisiana > Orleans Parish
- New Orleans (0.04)
- California > Santa Clara County
- Sunnyvale (0.04)
- Europe
- Spain > Catalonia
- Barcelona Province > Barcelona (0.04)
- Italy > Calabria
- Catanzaro Province > Catanzaro (0.04)
- Spain > Catalonia
- Asia > Myanmar
- Tanintharyi Region > Dawei (0.04)
- North America > United States
- Genre:
- Research Report > New Finding (0.88)
- Technology: