ref-hallucination-arena

You detect LLM reference hallucinations by verifying generated citations against major academic databases like Crossref, PubMed, arXiv, and DBLP. This quantifies fabrication rates and measures per-field citation accuracy for AI-generated literature reviews.

You can verify LLM citation accuracy against Crossref, PubMed, arXiv, and DBLP. Validating cited papers across these databases measures discipline-specific compliance and detects fabricated academic references in generated text.

Benchmark LLM literature reviews by validating recommended papers against Crossref, PubMed, arXiv, and DBLP. This quantifies hallucination rates and measures per-field citation accuracy, including title, author, year, and DOI match rates.

Yes, LLM hallucination detection supports tool-augmented modes for enhanced accuracy. This approach evaluates LLM reference recommendation capabilities by verifying cited papers against Crossref, PubMed, arXiv, and DBLP to ensure citation reliability.

Evaluating LLM reference recommendation capabilities uses metrics including hallucination rate, per-field accuracy for title, author, year, and DOI, plus discipline-specific compliance measured against verified academic databases like Crossref and PubMed.

Yes, matplotlib is required to benchmark LLM reference accuracy. This dependency supports the visualization of hallucination rates and per-field citation accuracy metrics when evaluating AI-generated academic references.