Comprehensive benchmark synthesizing the theoretical causes of factual hallucinations in LLMs. Unpacks internal activation entropy probing, retrieval-augmented grounding (RAG), self-consistency cross-examination, and decoding-time probability calibration mechanisms.
@article{sajjadi2026hallucination,
title = {A Survey on Hallucination in Large Language Models: Definitions, Detection, and Mitigation},
author = {Sajjadi Mohammadabadi, Seyed Mahmoud and Kara, Burak Cem and Eyupoglu, Can and Karakus, Oktay},
journal = {Preprints},
year = {2026},
doi = {10.20944/preprints202510.0000.v1}
}