Enhancing Answer Attribution for Faithful Text Generation with Large Language Models

TL;DR

This work performs a case study analyzing the effectiveness of existing answer attribution methods, with a focus on subtasks of answer segmentation and evidence retrieval, and proposes new methods for producing more independent and contextualized claims for better retrieval and attribution.

Abstract

The increasing popularity of Large Language Models (LLMs) in recent years has changed the way users interact with and pose questions to AI-based conversational systems. An essential aspect for increasing the trustworthiness of generated LLM answers is the ability to trace the individual claims from responses back to relevant sources that support them, the process known as answer attribution. While recent work has started exploring the task of answer attribution in LLMs, some challenges still remain. In this work, we first perform a case study analyzing the effectiveness of existing answer attribution methods, with a focus on subtasks of answer segmentation and evidence retrieval. Based on the observed shortcomings, we propose new methods for producing more independent and contextualized claims for better retrieval and attribution. The new methods are evaluated and shown to improve the performance of answer attribution components. We end with a discussion and outline of future directions for the task.

Figures (3)

• Figure 1: The complete answer attribution process (in the Post-Hoc-Retrieval setup)
• Figure 2: Statistics of contextualization of the 290 created claims by GPT3.5 and GPT4, evaluated by GPT4
• Figure 3: Visualization of the factuality evaluation statistics for the four different systems