LLMBridge: An LLM Pipeline for End-to-end Referential Bridging Resolution in English

2026-05-27

AI Summary

This work addresses the end-to-end bridging anaphora resolution task in English—identifying implicit bridging anaphoric expressions in text and linking them to their antecedents—by proposing a novel approach based on large language models (LLMs). It is the first to systematically leverage the natural language inference capabilities of LLMs for bridging resolution, integrating heuristic preprocessing and postprocessing strategies to achieve fully end-to-end bridging relation identification without manual feature engineering. Evaluated on three standard datasets—ISNotes, BASHI, and GUMBridge—the proposed method significantly outperforms existing state-of-the-art systems under both end-to-end and gold bridging anaphora settings, demonstrating its effectiveness and strong generalization capacity.

Abstract

In this paper, we introduce LLMBridge, a new LLM based system for the task of end-to-end referential bridging resolution in English. Our bridging resolution pipeline combines heuristic pre/post-processing with the natural language inference ability that comes from LLMs. We evaluate our bridging resolution pipeline on 3 datasets which have been used for referential bridging resolution evaluation in English: ISNotes, BASHI, and GUMBridge. Comparison to previous bridging resolution systems shows that the performance of LLMBridge surpasses previous state-of-the-art (SoTA) systems for all 3 datasets in the challenging End-to-end Evaluation Setting, as well as the Basic Bridging Resolution Evaluation Setting (gold bridging anaphor given). We also conduct a thorough error analysis of the LLMBridge performance, examining what varieties of bridging remain difficult for LLM based systems to identify. With this paper, we release the code for the LLMBridge pipeline.

Problem

• referential bridging resolution
• end-to-end
• anaphora
• natural language inference
• LLM

Innovation

• LLM-based bridging resolution
• end-to-end referential bridging
• natural language inference
• heuristic pre/post-processing
• state-of-the-art performance