Are Emotion and Rhetoric Neurons in LLM? Neuron Recognition and Adaptive Masking for Emotion-Rhetoric Prediction Steering

2026-04-19

AI Summary

Existing approaches struggle to achieve fine-grained, neuron-level control over emotional and rhetorical expressions in large language models, and conventional masking techniques lack robust causal validation. This work presents the first joint investigation into the neuronal representation mechanisms of six emotion categories and four core rhetorical devices, along with their intrinsic interrelationships. We propose a neuron identification framework integrating multidimensional filtering and introduce an adaptive masking strategy comprising dynamic filtering, attenuation masking, and feedback-based optimization to enable causally verifiable, targeted intervention. Experiments across five standard datasets demonstrate the method’s effectiveness, significantly enhancing performance on emotion-related tasks and successfully inducing non-target syntactic constructions. Our approach establishes a novel paradigm for fine-grained modulation of emotional and rhetorical expression in large language models.

Technology Category

• Natural Language Processing: (Large) Language Models
• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Search and Retrieval-Augmented AI: Large language models for search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Accurate comprehension and controllable generation of emotion and rhetoric are pivotal for enhancing the reasoning capabilities of large language models (LLMs). Existing studies mostly rely on external optimizations, lacking in-depth exploration of internal representation mechanisms, thus failing to achieve fine-grained steering at the neuron level. A handful of works on neurons are confined to emotions, neglecting rhetoric neurons and their intrinsic connections. Traditional neuron masking also exhibits counterintuitive phenomena, making reliable verification of neuron functionality infeasible. To address these issues, we systematically investigate the neurons representation mechanisms and inherent associations of 6 emotion categories and 4 core rhetorical devices. We propose a neuron identification framework that integrates multi-dimensional screening, and design an adaptive masking method incorporating dynamic filtering, attenuation masking, and feedback optimization, enabling reliable causal validation of neuron functionality.Through neuron regulation, we achieve directed induction of non-target sentences and enhancement of emotion tasks via rhetoric neurons. Experiments on 5 commonly used datasets validate the effectiveness of our method, providing a novel paradigm for the fine-grained steering of emotion and rhetoric expressions in LLMs.

Problem

• emotion neurons
• rhetoric neurons
• neuron masking
• fine-grained steering
• large language models

Innovation

• neuron recognition
• adaptive masking
• emotion-rhetoric steering
• large language models
• causal validation