Skip to content

Steering Geometry: Validating Human Value Geometry in LLM Steering Space

Sep 2026 · Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing · Vol abs/2609.06289, pp. 57010-57033 · 2 citations · 45 references
Computer Science

TL;DR

It is found that distribution-driven methods recover human value topologies aligned with theoretical predictions, and better geometric alignment also leads to more human-consistent transfer across values: steering one value correctly lifts compatible values and suppresses opposing ones.

Abstract

As large language models (LLMs) are increasingly deployed in alignment-sensitive contexts, activation steering has emerged as a lightweight, inference-time alternative to fine-tuning methods (e.g., RLHF, DPO) for behavioral control. However, existing work typically validates steering on isolated behaviors, leaving it unclear whether steering vectors encode coherent semantic structure or merely exploit behavior-specific shortcuts. We investigate whether the latent geometry of LLM steering vectors reflects theory-specified structure in human values and morality. Using Schwartz's Theory of Basic Human Values as our primary fine-grained framework, we introduce a 26K-sample benchmark covering 20 human values and analyze distribution-driven methods (e.g., CAA, SphericalSteer, ODESteer) and behavior-centric approaches (e.g., COLD-Steer, BiPO) across diverse model families and sizes. We find that distribution-driven methods recover human value topologies aligned with theoretical predictions (Spearman $\rho$ up to 0.51, $p<10^{-13}$). In contrast, behavior-centric methods achieve comparable steering performance but show little correlation with the expected value geometry. Geometric fidelity improves with model scale but drops after instruction tuning. Finally, better geometric alignment also leads to more human-consistent transfer across values: steering one value correctly lifts compatible values and suppresses opposing ones. Code and data are available at: https://github.com/DeepRCL/Steering_Geometry.

Read PDF

Similar papers

Preprint Nov 2026

Towards Multi-Value Alignment via Value Localization in LLMs

Aligning large language models (LLMs) with multiple, potentially conflicting human values remains challenging, as existing multi-objective alignment methods often require costly fine-tuning and provide limited insight into the internal mechanisms underlying value-related behaviors. In this work, we investigate multi-va...

Xue-Qi Ma, Yan-Bei Jiang, Xingjun Ma et al. · 0 citations
Preprint Nov 2026

Sweep Less, Steer More: Using Activation Steering Geometry for Reduced Hyperparameter Search

Representation steering methods like difference-in-means (DIM) offer a lightweight approach to controlling LLM behavior, but their effectiveness varies widely across concepts, datasets, and layers, making principled deployment costly without exhaustive hyperparameter sweeps. We build on the discriminability index $d'$,...

N. Hundia, S. Agrawal, Navita Goyal et al. · 0 citations
Preprint Nov 2026

TopoSteer: Topological Layer Selection for Activation Steering in LLMs

Activation steering controls large language models at inference time by adding a vector to their internal representations. Most work asks how to build that vector, and far less attention has gone to where to add it, even though steering at a poorly chosen layer can be worse than not steering at all. Prior selectors cho...

Tuc Nguyen, Marcos Gutiérrez-del-Pozo, Eduardo Paluzo-Hidalgo et al. · 1 citation

Minimally Invasive Steering of Language Models

Minimally Invasive Steering Vector Optimization (MISVO), which penalizes interventions using the local KL geometry of the induced token distribution, achieves the highest mean reward in six of seven model--task settings, with diversity and coherence scores close to those of Best-of-N.

Taha Entesari, Jingyu (Jack) Zhang, Daniel Khashabi et al. · 0 citations

Adaptive Multi-Value Control in LLMs via Causal Activation Steering

Large language models (LLMs) are increasingly deployed in settings where responses must reflect multiple, potentially interacting social norms and human values. Activation steering offers a lightweight alternative to training-based alignment by modifying internal activations at inference time. However, prior human-valu...

Payel Bhattacharjee, Ravi Tandon · 1 citation
#artificial intelligence Preprint Oct 2026

On the Steering Dimensionality of Refusal in Language Models

Existing activation steering methods often assume that a high-level concept can be mediated by a single steering direction. To support this, two complementary interventions should be achieved: additive steering should induce the target behavior, while directional ablation should suppress it. Yet behaviors may occupy ri...

Han-Gong Wang, Erik Miehling, Dennis Wei et al. · 1 citation

Related blog posts

MIT News · Artificial Intelligence Nov 17, 2026

Learning the language of machine learning

By studying the computational foundations of language learning, Associate Professor Jacob Andreas builds intelligent systems that can communicate effectively with humans.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.