arXiv Machine Learning September 11, 2026 By Xuan Cuong Ngo, Hao Vo, Ngan Le

GEOSTEER: Geodesic Optimization for Activation Steering in Large Language Models

Read the original on arXiv Machine Learning →

GeoSteer introduces a geometry-aware, optimization-based approach to norm-preserving activation steering in large language models. By formulating steering as a Riemannian optimization problem, it updates activations through a sequence of small geodesic steps guided by a learned nonlinear objective, avoiding fixed steering directions. Experiments on TruthfulQA, RealToxicityPrompts, and UltraFeedback show that GeoSteer consistently outperforms existing activation steering baselines, offering smoother, more stable, and more consistent steering behavior.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

arXiv AI

Jun 9

Riemannian-Manifold Steering: Geometry-Aware Generative Autoencoders for Label-Free Steering

arXiv:2605. 24942v2 Announce Type: replace-cross Abstract: Steering a language model - intervening on its internal activations to change downstream behaviour - has recently expanded beyond linear interpolation to nonlinear methods such as angular and kernelized steering, which define intervention transformations without learning an explicit geometry over paths in activation space.

By Narmeen Oozeer, Shivam Raval, Philip Quirke, Manikandan Ravikiran, Jeff Phillips, Shriyash Upadhyay, Amirali Abdullah

llms benchmarks

arXiv Machine Learning

Jun 9

Beyond Linear Activation Steering: Invertible Latent Transformations for Controlling LLM Behavior

arXiv:2606. 08454v1 Announce Type: new Abstract: Activation steering provides a lightweight inference-time mechanism for controlling large language models (LLMs) by modifying their internal activation vectors toward desired behaviors.

By Tuc Nguyen, Thai Le

llms diffusion benchmarks safety

arXiv AI

Jul 14

Conditional Optimal Bridge for Riemannian Activation Steering

arXiv:2607. 10517v1 Announce Type: cross Abstract: Activation steering offers a lightweight alternative to fine-tuning for controlling large language models at inference time.

By Seyed Arshan Dalili, Ajay Narayanan Sridhar, Vijaykrishnan Narayanan, Mehrdad Mahdavi

llms fine-tuning safety

arXiv Computation and Language

Sep 10

RiLM: Parameter-Efficient Language Modeling via Geodesic Decoding

arXiv:2609.10305v1 Announce Type: new Abstract: Language models under one million parameters matter for edge deployment, domain adaptation, and reproducible research, yet a two-layer LSTM or Transfor...

By Fang Li

llms rag fine-tuning efficiency

arXiv Computation and Language

Sep 22

Steering Vector Fields for Context-Aware Inference-Time Control in Large Language Models

arXiv:2602.01654v2 Announce Type: replace Abstract: Steering vectors (SVs) offer a lightweight way to control large language models (LLMs) at inference time by shifting hidden activations, providing...

By Jiaqian Li, Yanshu Li, Kuan-Hao Huang

llms fine-tuning

arXiv AI

Jun 8

A Geometric Account of Activation Steering through Angle-Norm Decomposition

arXiv:2606. 06735v1 Announce Type: new Abstract: Linear activation steering has gained popularity as a simple and empirically effective way to control language model behavior.

By Georgii Aparin, Tatiana Gaintseva

llms safety

One email a morning, machine-written

Related stories

Riemannian-Manifold Steering: Geometry-Aware Generative Autoencoders for Label-Free Steering

Beyond Linear Activation Steering: Invertible Latent Transformations for Controlling LLM Behavior

Conditional Optimal Bridge for Riemannian Activation Steering

RiLM: Parameter-Efficient Language Modeling via Geodesic Decoding

Steering Vector Fields for Context-Aware Inference-Time Control in Large Language Models

A Geometric Account of Activation Steering through Angle-Norm Decomposition