H-Neurons: 关于大语言模型中与幻觉相关的神经元的存在、影响及其起源
H-Neurons: On the Existence, Impact, and Origin of Hallucination-Associated Neurons in LLMs
AI总结：
本文研究了大语言模型中与幻觉相关的神经元的存在、影响及起源，揭示了这些神经元在预测幻觉和因果行为中的作用，为提升模型可靠性提供新视角。
摘要：
大型语言模型（LLMs）经常生成幻觉——看似合理但事实错误的输出——这削弱了它们的可靠性。虽然先前的研究从宏观视角如训练数据和目标来考察幻觉，但底层的神经元级机制仍鲜有探索。在本文中，我们从三个方面系统地研究了LLMs中的幻觉相关神经元（H-Neurons）：识别、行为影响和起源。在识别方面，我们证明了一小部分神经元（不到总神经元的0.1%）能够可靠地预测幻觉的发生，且在不同场景中具有良好的泛化能力。在行为影响方面，受控干预表明这些神经元与过度顺从行为存在因果联系。在起源方面，我们追溯这些神经元到预训练基础模型，并发现这些神经元在幻觉检测中仍然具有预测性，表明它们在预训练过程中出现。我们的发现将宏观行为模式与微观神经机制联系起来，为开发更可靠的LLMs提供了见解。
英文摘要：
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.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
