Image de couverture de livre

On the Non-Identifiability of Steering Vectors in Large Language Models

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Venkatesh, Sohan, Kurapath, Ashish Mahendran
Format:	Preprint
Publié:	2026
Sujets:	Machine Learning Artificial Intelligence
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Large Language Models are Algorithmically Blind
par: Venkatesh, Sohan, et autres
Publié: (2026)

Understanding Reasoning in Thinking Language Models via Steering Vectors
par: Venhoff, Constantin, et autres
Publié: (2025)

Compositional Steering of Large Language Models with Steering Tokens
par: Radevski, Gorjan, et autres
Publié: (2026)

Steering Large Language Model Activations in Sparse Spaces
par: Bayat, Reza, et autres
Publié: (2025)

Dialz: A Python Toolkit for Steering Vectors
par: Siddique, Zara, et autres
Publié: (2025)

To Steer or Not to Steer? Mechanistic Error Reduction with Abstention for Language Models
par: Hedström, Anna, et autres
Publié: (2025)

Generating Realistic Tabular Data with Large Language Models
par: Nguyen, Dang, et autres
Publié: (2024)

Discovering Reinforcement Learning Interfaces with Large Language Models
par: Jaswal, Akshat Singh, et autres
Publié: (2026)

Understanding Unreliability of Steering Vectors in Language Models: Geometric Predictors and the Limits of Linear Approximations
par: Braun, Joschka
Publié: (2026)

Adaptive Acquisition Selection for Bayesian Optimization with Large Language Models
par: Ngo, Giang, et autres
Publié: (2026)

Steering Large Language Models for Machine Translation Personalization
par: Scalena, Daniel, et autres
Publié: (2025)

Revis: Sparse Latent Steering to Mitigate Object Hallucination in Large Vision-Language Models
par: Wu, Jialin, et autres
Publié: (2026)

SPaCe: Unlocking Sample-Efficient Large Language Models Training With Self-Pace Curriculum Learning
par: Do, Dai, et autres
Publié: (2025)

COLD-Steer: Steering Large Language Models via In-Context One-step Learning Dynamics
par: Sharma, Kartik, et autres
Publié: (2026)

Enhancing LLM Steering through Sparse Autoencoder-Based Vector Refinement
par: Wang, Anyi, et autres
Publié: (2025)

MLLMEraser: Achieving Test-Time Unlearning in Multimodal Large Language Models through Activation Steering
par: Ding, Chenlu, et autres
Publié: (2025)

Negligible in Size, Significant in Effect: On Scale Vectors in Large Language Models
par: Wang, Mingze, et autres
Publié: (2026)

FineSteer: A Unified Framework for Fine-Grained Inference-Time Steering in Large Language Models
par: Weng, Zixuan, et autres
Publié: (2026)

Large Language Models for Imbalanced Classification: Diversity makes the difference
par: Nguyen, Dang, et autres
Publié: (2025)

Steered LLM Activations are Non-Surjective
par: Mishra, Aayush, et autres
Publié: (2026)

One-shot Optimized Steering Vectors Mediate Safety-relevant Behaviors in LLMs
par: Dunefsky, Jacob, et autres
Publié: (2025)

Steer2Adapt: Dynamically Composing Steering Vectors Elicits Efficient Adaptation of LLMs
par: Han, Pengrui, et autres
Publié: (2026)

ContextFocus: Activation Steering for Contextual Faithfulness in Large Language Models
par: Anand, Nikhil, et autres
Publié: (2026)

Multi-property Steering of Large Language Models with Dynamic Activation Composition
par: Scalena, Daniel, et autres
Publié: (2024)

Interpretable Steering of Large Language Models with Feature Guided Activation Additions
par: Soo, Samuel, et autres
Publié: (2025)

Improving Steering Vectors by Targeting Sparse Autoencoder Features
par: Chalnev, Sviatoslav, et autres
Publié: (2024)

PiSSA: Principal Singular Values and Singular Vectors Adaptation of Large Language Models
par: Meng, Fanxu, et autres
Publié: (2024)

PCDVQ: Enhancing Vector Quantization for Large Language Models via Polar Coordinate Decoupling
par: Yue, Yuxuan, et autres
Publié: (2025)

MidSteer: Optimal Affine Framework for Steering Generative Models
par: Gaintseva, Tatiana, et autres
Publié: (2026)

Amnesia: Adversarial Semantic Layer Specific Activation Steering in Large Language Models
par: Raza, Ali, et autres
Publié: (2026)

Word Embeddings Are Steers for Language Models
par: Han, Chi, et autres
Publié: (2023)

Enhancing Large Language Models for Time-Series Forecasting via Vector-Injected In-Context Learning
par: Zhang, Jianqi, et autres
Publié: (2026)

Shifting Perspectives: Steering Vectors for Robust Bias Mitigation in LLMs
par: Siddique, Zara, et autres
Publié: (2025)

Endogenous Resistance to Activation Steering in Language Models
par: McKenzie, Alex, et autres
Publié: (2026)

Pruning Large Language Models by Identifying and Preserving Functional Networks
par: Liu, Yiheng, et autres
Publié: (2025)

Automating Steering for Safe Multimodal Large Language Models
par: Wu, Lyucheng, et autres
Publié: (2025)

Small Generalizable Prompt Predictive Models Can Steer Efficient RL Post-Training of Large Reasoning Models
par: Qu, Yun, et autres
Publié: (2026)

Model-Based Reinforcement Learning Under Confounding
par: Venkatesh, Nishanth, et autres
Publié: (2025)

On the Reasoning Abilities of Masked Diffusion Language Models
par: Svete, Anej, et autres
Publié: (2025)

KBVQ-MoE: KLT-guided SVD with Bias-Corrected Vector Quantization for MoE Large Language Models
par: Xu, Zukang, et autres
Publié: (2026)