Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.13359 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915861355298816 |
|---|---|
| author | Alagharu, Rishab Singh, Ishneet Sukhvinder Shamsudeen, Shaibi Wu, Zhen Panda, Ashwinee |
| author_facet | Alagharu, Rishab Singh, Ishneet Sukhvinder Shamsudeen, Shaibi Wu, Zhen Panda, Ashwinee |
| contents | Language models are commonly fine-tuned for safety alignment to refuse harmful prompts. One approach fine-tunes them to generate categorical refusal tokens that distinguish different refusal types before responding. In this work, we leverage a version of Llama 3 8B fine-tuned with these categorical refusal tokens to enable inference-time control over fine-grained refusal behavior, improving both safety and reliability. We show that refusal token fine-tuning induces separable, category-aligned directions in the residual stream, which we extract and use to construct categorical steering vectors with a lightweight probe that determines whether to steer toward or away from refusal during inference. In addition, we introduce a learned low-rank combination that mixes these category directions in a whitened, orthonormal steering basis, resulting in a single controllable intervention under activation-space anisotropy, and show that this intervention is transferable across same-architecture model variants without additional training. Across benchmarks, both categorical steering vectors and the low-rank combination consistently reduce over-refusals on benign prompts while increasing refusal rates on harmful prompts, highlighting their utility for multi-category refusal control. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_13359 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | From Refusal Tokens to Refusal Control: Discovering and Steering Category-Specific Refusal Directions Alagharu, Rishab Singh, Ishneet Sukhvinder Shamsudeen, Shaibi Wu, Zhen Panda, Ashwinee Artificial Intelligence Language models are commonly fine-tuned for safety alignment to refuse harmful prompts. One approach fine-tunes them to generate categorical refusal tokens that distinguish different refusal types before responding. In this work, we leverage a version of Llama 3 8B fine-tuned with these categorical refusal tokens to enable inference-time control over fine-grained refusal behavior, improving both safety and reliability. We show that refusal token fine-tuning induces separable, category-aligned directions in the residual stream, which we extract and use to construct categorical steering vectors with a lightweight probe that determines whether to steer toward or away from refusal during inference. In addition, we introduce a learned low-rank combination that mixes these category directions in a whitened, orthonormal steering basis, resulting in a single controllable intervention under activation-space anisotropy, and show that this intervention is transferable across same-architecture model variants without additional training. Across benchmarks, both categorical steering vectors and the low-rank combination consistently reduce over-refusals on benign prompts while increasing refusal rates on harmful prompts, highlighting their utility for multi-category refusal control. |
| title | From Refusal Tokens to Refusal Control: Discovering and Steering Category-Specific Refusal Directions |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2603.13359 |