When is Your LLM Steerable?

Hugging Face Daily Papers Papers

Summary

This paper introduces a method to predict activation steering effectiveness in language models from early decoding states using a Gradient Boosting Decision Trees (GBDT) classifier, enabling efficient steering strength optimization without full rollouts.

Activation steering offers a lightweight approach to control language models' behavior at inference time, but whether it succeeds or fails heavily depends on the prompt, concept, model, and steering configuration. Finding the regime and boundaries of successful steering typically requires expensive grid searches and post-hoc evaluation of full autoregressive rollouts. In this work, we investigate whether steerability can be predicted from the model's internal states at the beginning of the generation process, e.g., after generating the first few tokens, and how to leverage such a predictor to improve steering success rate. To this end, we first introduce ASTEER, a testbed including 1.4M steered generations, spanning 150 concepts with each steering success/failure labeled. Leveraging this testbed, we analyze the model's early decoding dynamics by extracting features that compare hidden states before and after steering across layers and initial decoding steps. These features help us understand how steering's effects propagate along layers and token positions, which provide key information for steerability prediction. We then train a Gradient Boosting Decision Trees (GBDT) classifier on these features to predict whether an intervention will under-steer, succeed, or over-steer without requiring full rollout. Our predictor achieves around 0.7 macro-F1 score on unseen concepts, demonstrating that early hidden states encode substantial, structured information about eventual steering efficacy. We further leverage this steerability predictor as guidance for steering strength searching, achieving near-optimal performance with a small fraction of decoding cost.
Original Article
View Cached Full Text

Cached at: 06/15/26, 09:06 AM

Paper page - When is Your LLM Steerable?

Source: https://huggingface.co/papers/2606.11599

Abstract

Activation steering effectiveness can be predicted from early decoding states using a GBDT classifier, enabling efficient steering strength optimization with reduced computational cost.

Activation steeringoffers a lightweight approach to controllanguage models’ behavior atinference time, but whether it succeeds or fails heavily depends on the prompt, concept, model, and steering configuration. Finding the regime and boundaries of successful steering typically requires expensivegrid searchesand post-hoc evaluation of fullautoregressive rollouts. In this work, we investigate whether steerability can be predicted from the model’s internal states at the beginning of the generation process, e.g., after generating the first few tokens, and how to leverage such a predictor to improve steering success rate. To this end, we first introduce ASTEER, a testbed including 1.4M steered generations, spanning 150 concepts with each steering success/failure labeled. Leveraging this testbed, we analyze the model’searly decoding dynamicsby extracting features that comparehidden statesbefore and after steering across layers and initial decoding steps. These features help us understand how steering’s effects propagate along layers and token positions, which provide key information forsteerability prediction. We then train aGradient Boosting Decision Trees(GBDT) classifier on these features to predict whether an intervention will under-steer, succeed, or over-steer without requiring full rollout. Our predictor achieves around 0.7 macro-F1 score on unseen concepts, demonstrating that earlyhidden statesencode substantial, structured information about eventual steering efficacy. We further leverage this steerability predictor as guidance forsteering strength searching, achieving near-optimal performance with a small fraction of decoding cost.

View arXiv pageView PDFGitHub3Add to collection

Get this paper in your agent:

hf papers read 2606\.11599

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2606.11599 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2606.11599 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2606.11599 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

When is Your LLM Steerable?

arXiv cs.CL

This paper investigates when activation steering succeeds or fails for LLMs by analyzing early decoding dynamics. The authors introduce ASTEER, a large testbed of steered generations, and train a GBDT classifier to predict steering outcomes from early hidden states, enabling efficient steering strength search.

Multi-Attribute Steering of Language Models via Targeted Intervention

arXiv cs.CL

MAT-Steer introduces a novel inference-time intervention framework for steering LLMs across multiple conflicting attributes by learning sparse, orthogonal steering vectors that selectively target tokens relevant to each attribute, achieving gains in QA tasks and generative tasks over prior methods.

SALSA: Speech Aware LLM Adaptation via Learned Steering Activation Vectors

arXiv cs.CL

SALSA introduces a lightweight adaptation method for speech-aware LLMs that learns layer-wise steering vectors via supervised objective, achieving significant improvements (up to 46.8% relative) on out-of-domain speech benchmarks, and shows that steering the encoder layers is more effective than modifying the LLM backbone.

Steered LLM Activations are Non-Surjective

Hugging Face Daily Papers

This paper proves that activation steering in LLMs produces internal states that cannot be replicated by any textual prompt, establishing a formal separation between white-box steerability and black-box prompting.