Are Emotion and Rhetoric Neurons in LLM? Neuron Recognition and Adaptive Masking for Emotion-Rhetoric Prediction Steering
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914488601542656 |
|---|---|
| author | Zheng, Li Zhang, Xin He, Shuyi Li, Fei Teng, Chong Yang, Jiangming Ji, Donghong Li, Zhuang |
| author_facet | Zheng, Li Zhang, Xin He, Shuyi Li, Fei Teng, Chong Yang, Jiangming Ji, Donghong Li, Zhuang |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_17255 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Are Emotion and Rhetoric Neurons in LLM? Neuron Recognition and Adaptive Masking for Emotion-Rhetoric Prediction Steering Zheng, Li Zhang, Xin He, Shuyi Li, Fei Teng, Chong Yang, Jiangming Ji, Donghong Li, Zhuang Computation and Language 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. |
| title | Are Emotion and Rhetoric Neurons in LLM? Neuron Recognition and Adaptive Masking for Emotion-Rhetoric Prediction Steering |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2604.17255 |