Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
October 13, 2025Open Access

Improving LLMs' Learning for Coreference Resolution

View Full Paper
Ask AI
Bookmark
Share

Authors

YGYujian GanYLYuan LiangYLYanni Lin

Discussion

Loading...

Member takes

Overview

Novel techniques improve coreference resolution and tackle hallucination in large language models.

Key Points

  • Reversed training significantly enhances the performance of the question-answering template method.
  • Iterative document generation effectively eliminates hallucinations, boosting overall coreference resolution.
  • Investigated limitations of LLM-based approaches and proposed integrated novel techniques for improvement.
  • Enhanced models offer a robust solution to the persistent challenges in coreference resolution tasks.

Cite This Study

Gan et al. (2025) studied this question.

synapsesocial.com/papers/68ecfebf950606aabec094cdhttps://doi.org/10.48550/arxiv.2509.11466
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Findings of the Fourth Shared Task on Multilingual Coreference Resolution: Can LLMs Dethrone Traditional Approaches?2025
  2. 2MrRank: Improving Question Answering Retrieval System through Multi-Result Ranking Model2024
  3. 3Rejection Improves Reliability: Training LLMs to Refuse Unknown Questions Using RL from Knowledge Feedback2024 · 2 citations
  4. 4Hallucination Detection: Robustly Discerning Reliable Answers in Large Language Models2024
  5. 5Improving the Reliability of LLMs: Combining CoT, RAG, Self-Consistency, and Self-Verification2025 · 2 citations