Source or It Didn't Happen: A Multi-Agent Framework for Citation Hallucination Detection

TL;DR

This paper introduces CiteTracer, a multi-agent framework for detecting citation hallucinations in scientific writing by leveraging structured taxonomy, evidence retrieval, and class-specific adjudication.

Contribution

It presents a novel taxonomy-aligned adjudication framework and a multi-agent detector that significantly improves citation hallucination detection accuracy.

Findings

CiteTracer achieves 97.1% accuracy on synthetic benchmarks.

Detects 97.1% of fabrications in real-world citations.

Outperforms existing detectors in class-level F1 scores.

Abstract

Large language models are increasingly used in scientific writing, yet they can fabricate citation-shaped references that appear plausible but fail bibliographic verification. Existing detectors often reduce verification to binary found/not-found decisions and rely on brittle parsing or incomplete retrieval, offering little field-level signal to auditors. We reframe citation hallucination detection as taxonomy-aligned field-level adjudication and introduce a 12-code taxonomy spanning Real, Potential, and Hallucinated citations. Based on this taxonomy, we build CiteTracer, a cascading multi-agent detector that extracts structured citations from PDF and BibTeX, retrieves evidence through cache lookup, URL fetch, scholar connectors, and web search, applies deterministic field matching, and routes ambiguous cases to class-specialist judgers. We release a benchmark of 2,450 synthetic citations built from real seeds with controlled LLM mutations, paired with 957 real-world fabricated citations drawn from ICLR 2026 and an anonymous conference desk-rejected submissions. CiteTracer reaches 97.1% accuracy on the synthetic benchmark, with class-level F1 scores of 97.0, 95.8, and 98.5 for Real, Potential, and Hallucinated, respectively, and detects 97.1% of fabrications on the real-world set without abstaining. Code: https://github.com/aaFrostnova/CiteTracer.