Enhancing LLM Steering through Sparse Autoencoder-Based Vector Refinement
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866918153559212032 |
|---|---|
| author | Wang, Anyi Wu, Xuansheng Shu, Dong Ma, Yunpu Liu, Ninghao |
| author_facet | Wang, Anyi Wu, Xuansheng Shu, Dong Ma, Yunpu Liu, Ninghao |
| contents | Steering has emerged as a promising approach in controlling large language models (LLMs) without modifying model parameters. However, most existing steering methods rely on large-scale datasets to learn clear behavioral information, which limits their applicability in many real-world scenarios. The steering vectors extracted from small dataset often contain task-irrelevant noising features, which degrades their effectiveness. To refine the steering vectors learned from limited data, we introduce Refinement of Steering Vector via Sparse Autoencoder (SAE-RSV) that leverages SAEs to semantically denoise and augment the steering vectors. In our framework, we first remove task-irrelevant features according to their semantics provided by SAEs, and then enrich task-relevant features missing from the small dataset through their semantic similarity to the identified relevant features. Extensive experiments demonstrate that the proposed SAE-RSV substantially outperforms all the baseline methods including supervised fine-tuning. Our findings show that effective steering vector can be constructed from limited training data by refining the original steering vector through SAEs. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_23799 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Enhancing LLM Steering through Sparse Autoencoder-Based Vector Refinement Wang, Anyi Wu, Xuansheng Shu, Dong Ma, Yunpu Liu, Ninghao Machine Learning Artificial Intelligence Steering has emerged as a promising approach in controlling large language models (LLMs) without modifying model parameters. However, most existing steering methods rely on large-scale datasets to learn clear behavioral information, which limits their applicability in many real-world scenarios. The steering vectors extracted from small dataset often contain task-irrelevant noising features, which degrades their effectiveness. To refine the steering vectors learned from limited data, we introduce Refinement of Steering Vector via Sparse Autoencoder (SAE-RSV) that leverages SAEs to semantically denoise and augment the steering vectors. In our framework, we first remove task-irrelevant features according to their semantics provided by SAEs, and then enrich task-relevant features missing from the small dataset through their semantic similarity to the identified relevant features. Extensive experiments demonstrate that the proposed SAE-RSV substantially outperforms all the baseline methods including supervised fine-tuning. Our findings show that effective steering vector can be constructed from limited training data by refining the original steering vector through SAEs. |
| title | Enhancing LLM Steering through Sparse Autoencoder-Based Vector Refinement |
| topic | Machine Learning Artificial Intelligence |
| url | https://arxiv.org/abs/2509.23799 |