Image de couverture de livre

ExpertLens: Activation steering features are highly interpretable

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Fedzechkina, Masha, Gualdoni, Eleonora, Williamson, Sinead, Metcalf, Katherine, Seto, Skyler, Theobald, Barry-John
Format:	Preprint
Publié:	2025
Sujets:	Computation and Language Artificial Intelligence
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Steering into New Embedding Spaces: Analyzing Cross-Lingual Alignment Induced by Model Interventions in Multilingual Language Models
par: Sundar, Anirudh, et autres
Publié: (2025)

What do your logits know? (The answer may surprise you!)
par: Fedzechkina, Masha, et autres
Publié: (2026)

LinEAS: End-to-end Learning of Activation Steering with a Distributional Loss
par: Rodriguez, Pau, et autres
Publié: (2025)

Investigating Intersectional Bias in Large Language Models using Confidence Disparities in Coreference Resolution
par: Khan, Falaah Arif, et autres
Publié: (2025)

Discriminating Form and Meaning in Multilingual Models with Minimal-Pair ABX Tasks
par: de Seyssel, Maureen, et autres
Publié: (2025)

PREDICT: Preference Reasoning by Evaluating Decomposed preferences Inferred from Candidate Trajectories
par: Aroca-Ouellette, Stephane, et autres
Publié: (2024)

Whispering Experts: Neural Interventions for Toxicity Mitigation in Language Models
par: Suau, Xavier, et autres
Publié: (2024)

Can sparse autoencoders be used to decompose and interpret steering vectors?
par: Mayne, Harry, et autres
Publié: (2024)

Sample-Efficient Preference-based Reinforcement Learning with Dynamics Aware Rewards
par: Metcalf, Katherine, et autres
Publié: (2024)

On the Limited Generalization Capability of the Implicit Reward Model Induced by Direct Preference Optimization
par: Lin, Yong, et autres
Publié: (2024)

DSO: Direct Steering Optimization for Bias Mitigation
par: Paes, Lucas Monteiro, et autres
Publié: (2025)

Is Your Model Fairly Certain? Uncertainty-Aware Fairness Evaluation for LLMs
par: Wang, Yinong Oliver, et autres
Publié: (2025)

Why do objects have many names? A study on word informativeness in language use and lexical systems
par: Gualdoni, Eleonora, et autres
Publié: (2024)

Analyzing the Effect of Linguistic Similarity on Cross-Lingual Transfer: Tasks and Experimental Setups Matter
par: Blaschke, Verena, et autres
Publié: (2025)

Aligning LLMs by Predicting Preferences from User Writing Samples
par: Aroca-Ouellette, Stéphane, et autres
Publié: (2025)

RAGBench: Explainable Benchmark for Retrieval-Augmented Generation Systems
par: Friel, Robert, et autres
Publié: (2024)

Evaluating Gender Bias Transfer between Pre-trained and Prompt-Adapted Language Models
par: Mackraz, Natalie, et autres
Publié: (2024)

Enabling Precise Topic Alignment in Large Language Models Via Sparse Autoencoders
par: Joshi, Ananya, et autres
Publié: (2025)

Exploring and steering the moral compass of Large Language Models
par: Tlaie, Alejandro
Publié: (2024)

Luna: An Evaluation Foundation Model to Catch Language Model Hallucinations with High Accuracy and Low Cost
par: Belyi, Masha, et autres
Publié: (2024)

Reasoning's Razor: Reasoning Improves Accuracy but Can Hurt Recall at Critical Operating Points in Safety and Hallucination Detection
par: Chegini, Atoosa, et autres
Publié: (2025)

Revisiting Uncertainty Quantification Evaluation in Language Models: Spurious Interactions with Response Length Bias Results
par: Santilli, Andrea, et autres
Publié: (2025)

SelfReflect: Can LLMs Communicate Their Internal Answer Distribution?
par: Kirchhof, Michael, et autres
Publié: (2025)

Navigating through the hidden embedding space: steering LLMs to improve mental health assessment
par: Ravenda, Federico, et autres
Publié: (2025)

Toward universal steering and monitoring of AI models
par: Beaglehole, Daniel, et autres
Publié: (2025)

Alloc-MoE: Budget-Aware Expert Activation Allocation for Efficient Mixture-of-Experts Inference
par: Liu, Baihui, et autres
Publié: (2026)

BED-LLM: Intelligent Information Gathering with LLMs and Bayesian Experimental Design
par: Choudhury, Deepro, et autres
Publié: (2025)

Closing the Gap Between Text and Speech Understanding in LLMs
par: Cuervo, Santiago, et autres
Publié: (2025)

StructLens: A Structural Lens for Language Models via Maximum Spanning Trees
par: Sakajo, Haruki, et autres
Publié: (2026)

Localizing Persona Representations in LLMs
par: Cintas, Celia, et autres
Publié: (2025)

Automating Historical Insight Extraction from Large-Scale Newspaper Archives via Neural Topic Modeling
par: Murugaraj, Keerthana, et autres
Publié: (2025)

Soup-of-Experts: Pretraining Specialist Models via Parameters Averaging
par: Ablin, Pierre, et autres
Publié: (2025)

HalluLens: LLM Hallucination Benchmark
par: Bang, Yejin, et autres
Publié: (2025)

FaithLens: Detecting and Explaining Faithfulness Hallucination
par: Si, Shuzheng, et autres
Publié: (2025)

Rectifying LLM Thought from Lens of Optimization
par: Liu, Junnan, et autres
Publié: (2025)

ExpertFlow: Efficient Mixture-of-Experts Inference via Predictive Expert Caching and Token Scheduling
par: He, Xin, et autres
Publié: (2024)

ExpertQA: Expert-Curated Questions and Attributed Answers
par: Malaviya, Chaitanya, et autres
Publié: (2023)

An Expert is Worth One Token: Synergizing Multiple Expert LLMs as Generalist via Expert Token Routing
par: Chai, Ziwei, et autres
Publié: (2024)

FMEA Builder: Expert Guided Text Generation for Equipment Maintenance
par: Lynch, Karol, et autres
Publié: (2024)

ExpertPrompting: Instructing Large Language Models to be Distinguished Experts
par: Xu, Benfeng, et autres
Publié: (2023)