- Rating
- 1
- Confidence
- 4
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
This paper proposes SteerCLR, an unsupervised method for jointly learning a bank of steering vectors on a frozen LLM. The method combines downstream activation-shift maximization with a contrastive objective that encourages consistent effects across prompts for the same vector and distinct effects across different vectors, without requiring behavioral labels during training.
Critical strengths
Unlike conventional methods such as CAA, which require predefined behavioral concepts and corresponding examples, SteerCLR jointly optimizes a bank of vectors using unlabeled prompts. The contrastive objective operates on the downstream activation changes induced by steering, which is a reasonable design choice, since geometrically distinct vectors do not necessarily produce distinct downstream effects.
Critical weaknesses
- The main evaluation measures steering effects along seven predefined behavioral axes. Although SteerCLR achieves higher functional effective rank and lower behavioral collision rates, these metrics only describe how the learned vectors respond along those seven axes. They do not establish how many distinct, interpretable concepts the method actually discovers. The qualitative examples also provide limited support for the central discovery claim. Most demonstrated behaviors, such as switching languages, adopting a pirate persona, or changing writing styles, are familiar concepts that can already be controlled using methods like Activation Addition and CAA.
- SteerCLR maximizes the magnitude of downstream activation shifts and encourages separation between their directions. However, large activation changes do not necessarily imply effective steering. They may also introduce undesirable changes in generation quality or unrelated behaviors. Likewise, distinct activation shifts do not necessarily correspond to distinct semantic effects. The main evaluation focuses on the distribution of steering effects across seven behavioral axes, measured using multiple-choice answer probabilities. It does not directly establish whether the learned vectors provide stronger, more consistent, or more reliable behavioral control on natural generation tasks.
Questions for authors
- This paper uses different bank sizes, model architectures, and training prompt distributions for the results. It does not explain these particular choices. Do the authors have results showing how the number of vectors affects behavioral coverage or redundancy? In particular, does increasing the bank size discover additional concepts, subdivide existing concepts, or primarily introduce redundant vectors?
- The qualitative analysis relies on Qwen2.5-72B-Instruct to assign labels based on a small number of paired generations. Have the authors examined whether different labeling models or different prompt subsets produce consistent labels for the same vector?
- For concepts identified by the labeling model, such as language switching or refusal, could the authors compare SteerCLR vectors against vectors obtained through CAA on the same base model? This would help clarify what unsupervised discovery provides beyond established methods.
- The abstract states that discovering steering vectors with distinct, interpretable effects remains difficult. However, existing approaches can construct effective steering vectors when the target concepts and corresponding examples are specified. The difficulty addressed here appears to be discovering diverse behavioral directions without predefined concepts, rather than constructing steering vectors in general. I suggest making this distinction more explicit.