Copertina

Whispering Experts: Neural Interventions for Toxicity Mitigation in Language Models

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Suau, Xavier, Delobelle, Pieter, Metcalf, Katherine, Joulin, Armand, Apostoloff, Nicholas, Zappella, Luca, Rodríguez, Pau
Natura:	Preprint
Pubblicazione:	2024
Soggetti:	Computation and Language Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Controlling Language and Diffusion Models by Transporting Activations
di: Rodriguez, Pau, et al.
Pubblicazione: (2024)

Investigating Intersectional Bias in Large Language Models using Confidence Disparities in Coreference Resolution
di: Khan, Falaah Arif, et al.
Pubblicazione: (2025)

LinEAS: End-to-end Learning of Activation Steering with a Distributional Loss
di: Rodriguez, Pau, et al.
Pubblicazione: (2025)

HyperTransport: Amortized Conditioning of T2I Generative Models
di: Maiorca, Valentino, et al.
Pubblicazione: (2026)

Evaluating Gender Bias Transfer between Pre-trained and Prompt-Adapted Language Models
di: Mackraz, Natalie, et al.
Pubblicazione: (2024)

Is Your Model Fairly Certain? Uncertainty-Aware Fairness Evaluation for LLMs
di: Wang, Yinong Oliver, et al.
Pubblicazione: (2025)

GenCtrl -- A Formal Controllability Toolkit for Generative Models
di: Cheng, Emily, et al.
Pubblicazione: (2026)

Whispers that Shake Foundations: Analyzing and Mitigating False Premise Hallucinations in Large Language Models
di: Yuan, Hongbang, et al.
Pubblicazione: (2024)

ParaRNN: Unlocking Parallel Training of Nonlinear RNNs for Large Language Models
di: Danieli, Federico, et al.
Pubblicazione: (2025)

Steering into New Embedding Spaces: Analyzing Cross-Lingual Alignment Induced by Model Interventions in Multilingual Language Models
di: Sundar, Anirudh, et al.
Pubblicazione: (2025)

Dynamically Scaled Activation Steering
di: Ferrando, Alex, et al.
Pubblicazione: (2025)

ExpertLens: Activation steering features are highly interpretable
di: Fedzechkina, Masha, et al.
Pubblicazione: (2025)

From One to Many: Expanding the Scope of Toxicity Mitigation in Language Models
di: Pozzobon, Luiza, et al.
Pubblicazione: (2024)

Whisper Finetuning on Nepali Language
di: Rijal, Sanjay, et al.
Pubblicazione: (2024)

Fairness Dynamics During Training
di: Patel, Krishna, et al.
Pubblicazione: (2025)

GloSS over Toxicity: Understanding and Mitigating Toxicity in LLMs via Global Toxic Subspace
di: Duan, Zenghao, et al.
Pubblicazione: (2025)

CoRet: Improved Retriever for Code Editing
di: Fehr, Fabio, et al.
Pubblicazione: (2025)

Bias after Prompting: Persistent Discrimination in Large Language Models
di: Sivakumar, Nivedha, et al.
Pubblicazione: (2025)

Revisiting Uncertainty Quantification Evaluation in Language Models: Spurious Interactions with Response Length Bias Results
di: Santilli, Andrea, et al.
Pubblicazione: (2025)

How Toxic Can You Get? Search-based Toxicity Testing for Large Language Models
di: Corbo, Simone, et al.
Pubblicazione: (2025)

Time Sensitive Knowledge Editing through Efficient Finetuning
di: Ge, Xiou, et al.
Pubblicazione: (2024)

Large Language Model Adaptation for Financial Sentiment Analysis
di: Inserte, Pau Rodriguez, et al.
Pubblicazione: (2024)

ProbLog4Fairness: A Neurosymbolic Approach to Modeling and Mitigating Bias
di: Adriaensen, Rik, et al.
Pubblicazione: (2025)

DSO: Direct Steering Optimization for Bias Mitigation
di: Paes, Lucas Monteiro, et al.
Pubblicazione: (2025)

Characterising Toxicity in Generative Large Language Models
di: Zhang, Zhiyao, et al.
Pubblicazione: (2026)

Realistic Evaluation of Toxicity in Large Language Models
di: Luong, Tinh Son, et al.
Pubblicazione: (2024)

Preference Tuning For Toxicity Mitigation Generalizes Across Languages
di: Li, Xiaochen, et al.
Pubblicazione: (2024)

Following the Whispers of Values: Unraveling Neural Mechanisms Behind Value-Oriented Behaviors in LLMs
di: Hu, Ling, et al.
Pubblicazione: (2025)

Interpreting CLIP: Insights on the Robustness to ImageNet Distribution Shifts
di: Crabbé, Jonathan, et al.
Pubblicazione: (2023)

ExpertPrompting: Instructing Large Language Models to be Distinguished Experts
di: Xu, Benfeng, et al.
Pubblicazione: (2023)

Focus on Your Question! Interpreting and Mitigating Toxic CoT Problems in Commonsense Reasoning
di: Li, Jiachun, et al.
Pubblicazione: (2024)

Large Language Models as Generalizable Policies for Embodied Tasks
di: Szot, Andrew, et al.
Pubblicazione: (2023)

Pre-Attention Expert Prediction and Prefetching for Mixture-of-Experts Large Language Models
di: Zhu, Shien, et al.
Pubblicazione: (2025)

Sparse Autoencoders are Capable LLM Jailbreak Mitigators
di: Assogba, Yannick, et al.
Pubblicazione: (2026)

Refining Positive and Toxic Samples for Dual Safety Self-Alignment of LLMs with Minimal Human Interventions
di: Xu, Jingxin, et al.
Pubblicazione: (2025)

Unplug and Play Language Models: Decomposing Experts in Language Models at Inference Time
di: Yang, Nakyeong, et al.
Pubblicazione: (2024)

Rethinking Toxicity Evaluation in Large Language Models: A Multi-Label Perspective
di: Kou, Zhiqiang, et al.
Pubblicazione: (2025)

Induction Head Toxicity Mechanistically Explains Repetition Curse in Large Language Models
di: Wang, Shuxun, et al.
Pubblicazione: (2025)

Pruning General Large Language Models into Customized Expert Models
di: Zhao, Yirao, et al.
Pubblicazione: (2025)

Mix of Experts Language Model for Named Entity Recognition
di: Chen, Xinwei, et al.
Pubblicazione: (2024)