H-Neurons: On the Existence, Impact, and Origin of Hallucination-Associated Neurons in LLMs

Large language models (LLMs) frequently generate hallucinations -- plausible but factually incorrect outputs -- undermining their reliability. While prior work has examined hallucinations from macroscopic perspectives such as training data and objectives, the underlying neuron-level mechanisms remain largely unexplored. In this paper, we conduct a systematic investigation into hallucination-associated neurons (H-Neurons) in LLMs from three perspectives: identification, behavioral impact, and origins. Regarding their identification, we demonstrate that a remarkably sparse subset of neurons (less than $0.1\%$ of total neurons) can reliably predict hallucination occurrences, with strong generalization across diverse scenarios. In terms of behavioral impact, controlled interventions reveal that these neurons are causally linked to over-compliance behaviors. Concerning their origins, we trace these neurons back to the pre-trained base models and find that these neurons remain predictive for hallucination detection, indicating they emerge during pre-training. Our findings bridge macroscopic behavioral patterns with microscopic neural mechanisms, offering insights for developing more reliable LLMs.

/via

Web & Technik #ki & assistenz

30.03.26 Bookmark

Florian Heinz

@hnz.io

Ich mach' was mit Kommunikation und Marketing, bin Podcaster auf Entzug, bleibe immer Hundepapa im Herzen und Tabler mit Leidenschaft. Hier landet alles, was mir Spaß macht: Texte, Fotos, Experimente. Göttingen ist mein Zuhause, das Netz mein Spielplatz.

Florian Heinz

Suchen & entdecken

Themen entdecken

Kommentar schreiben