DiscoTrace: Representing and Comparing Answering Strategies of Humans and LLMs in Information-Seeking Question Answering

TL;DR

DiscoTrace is a novel method that analyzes answer strategies in information-seeking QA, revealing differences between human communities and LLMs in rhetorical diversity and answer breadth.

Contribution

DiscoTrace introduces a new approach to identify and compare rhetorical strategies in answers, highlighting the lack of diversity in LLM responses compared to humans.

Findings

Humans show diverse answer construction strategies across communities.

LLMs lack rhetorical diversity and tend to address broader interpretations.

Findings can inform development of more context-aware LLM answerers.

Abstract

We introduce DiscoTrace, a method to identify the rhetorical strategies that answerers use when responding to information-seeking questions. DiscoTrace represents answers as a sequence of question-related discourse acts paired with interpretations of the original question, annotated on top of rhetorical structure theory parses. Applying DiscoTrace to answers from nine different human communities reveals that communities have diverse preferences for answer construction. In contrast, LLMs do not exhibit rhetorical diversity in their answers, even when prompted to mimic specific human community answering guidelines. LLMs also systematically opt for breadth, addressing interpretations of questions that human answerers choose not to address. Our findings can guide the development of pragmatic LLM answerers that consider a range of strategies informed by context in QA.