# A novel hallucination classification framework

arXiv: https://arxiv.org/abs/2510.05189 · HTML: https://mlsanity.com/paper/2510.05189/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 19 references that are in the map.
- First posted on arXiv 6 Oct 2025.

## Builds on (most cited first, 8 of 8)

- [UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction](https://mlsanity.com/paper/1802.03426/) (2018)
- [A Survey of Large Language Models](https://mlsanity.com/paper/2303.18223/) (2023)
- [A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions](https://mlsanity.com/paper/2311.05232/) (2023)
- [Can AI-Generated Text be Reliably Detected?](https://mlsanity.com/paper/2303.11156/) (2023)
- [Large Language Models for Information Retrieval: A Survey](https://mlsanity.com/paper/2308.07107/) (2023)
- [A Review on Language Models as Knowledge Bases](https://mlsanity.com/paper/2204.06031/) (2022)
- [Hallucinations in Large Multilingual Translation Models](https://mlsanity.com/paper/2303.16104/) (2023)
- [Unsupervised Real-Time Hallucination Detection based on the Internal States of Large Language Models](https://mlsanity.com/paper/2403.06448/) (2024)

## Abstract (arXiv)

This work introduces a novel methodology for the automatic detection of hallucinations generated during large language model (LLM) inference. The proposed approach is based on a systematic taxonomy and controlled reproduction of diverse hallucination types through prompt engineering. A dedicated hallucination dataset is subsequently mapped into a vector space using an embedding model and analyzed with unsupervised learning techniques in a reduced-dimensional representation of hallucinations with veridical responses. Quantitative evaluation of inter-centroid distances reveals a consistent correlation between the severity of informational distortion in hallucinations and their spatial divergence from the cluster of correct outputs. These findings provide theoretical and empirical evidence that even simple classification algorithms can reliably distinguish hallucinations from accurate responses within a single LLM, thereby offering a lightweight yet effective framework for improving model reliability.