Imagen de Portada

Combining Domain and Alignment Vectors to Achieve Better Knowledge-Safety Trade-offs in LLMs

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Thakkar, Megh, Fournier, Quentin, Riemer, Matthew, Chen, Pin-Yu, Zouaq, Amal, Das, Payel, Chandar, Sarath
Formato:	Preprint
Publicado:	2024
Materias:	Artificial Intelligence
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

A Deep Dive into the Trade-Offs of Parameter-Efficient Preference Alignment Techniques
por: Thakkar, Megh, et al.
Publicado: (2024)

LLMs Can't Play Hangman: On the Necessity of a Private Working Memory for Language Agents
por: Baldelli, Davide, et al.
Publicado: (2026)

Probabilistic Calibration Is a Trainable Capability in Language Models
por: Baldelli, Davide, et al.
Publicado: (2026)

Fundamental Safety-Capability Trade-offs in Fine-tuning Large Language Models
por: Chen, Pin-Yu, et al.
Publicado: (2025)

Exploring Quantization for Efficient Pre-Training of Transformer Language Models
por: Chitsaz, Kamran, et al.
Publicado: (2024)

Ontology-Constrained Generation of Domain-Specific Clinical Summaries
por: Mehenni, Gaya, et al.
Publicado: (2024)

Too Big to Fool: Resisting Deception in Language Models
por: Samsami, Mohammad Reza, et al.
Publicado: (2024)

Dialectics of Alignment: Harnessing Unsafe Knowledge for Dynamic Safety Routing
por: Hashemzadeh, Maryam, et al.
Publicado: (2026)

CADmium: Fine-Tuning Code Language Models for Text-Driven Sequential CAD Design
por: Govindarajan, Prashant, et al.
Publicado: (2025)

CoPeP: Benchmarking Continual Pretraining for Protein Language Models
por: Patil, Darshan, et al.
Publicado: (2026)

Manifold Metric: A Loss Landscape Approach for Predicting Model Performance
por: Malviya, Pranshu, et al.
Publicado: (2024)

NovoMolGen: Rethinking Molecular Language Model Pretraining
por: Chitsaz, Kamran, et al.
Publicado: (2025)

Revisiting Replay and Gradient Alignment for Continual Pre-Training of Large Language Models
por: Abbes, Istabrak, et al.
Publicado: (2025)

NeoBERT: A Next-Generation BERT
por: Breton, Lola Le, et al.
Publicado: (2025)

SEAL: Safety-enhanced Aligned LLM Fine-tuning via Bilevel Data Selection
por: Shen, Han, et al.
Publicado: (2024)

Small Encoders Can Rival Large Decoders in Detecting Groundedness
por: Abbes, Istabrak, et al.
Publicado: (2025)

What is the Best Process Model Representation? A Comparative Analysis for Process Modeling with Large Language Models
por: Brissard, Alexis, et al.
Publicado: (2025)

Towards Practical Tool Usage for Continually Learning LLMs
por: Huang, Jerry, et al.
Publicado: (2024)

The Unintended Trade-off of AI Alignment:Balancing Hallucination Mitigation and Safety in LLMs
por: Mahmoud, Omar, et al.
Publicado: (2025)

Reducing Hallucinations in Language Model-based SPARQL Query Generation Using Post-Generation Memory Retrieval
por: Sharma, Aditya, et al.
Publicado: (2025)

FRASE: Structured Representations for Generalizable SPARQL Query Generation
por: Diallo, Papa Abdou Karim Karou, et al.
Publicado: (2025)

Enhancing Frame Detection with Retrieval Augmented Generation
por: Diallo, Papa Abdou Karim Karou, et al.
Publicado: (2025)

Faithfulness Measurable Masked Language Models
por: Madsen, Andreas, et al.
Publicado: (2023)

Are self-explanations from Large Language Models faithful?
por: Madsen, Andreas, et al.
Publicado: (2024)

Patching LLM Like Software: A Lightweight Method for Improving Safety Policy in Large Language Models
por: Arif, Huzaifa, et al.
Publicado: (2025)

Reconstruction or Semantics? What Makes a Latent Space Useful for Robotic World Models
por: Nilaksh, et al.
Publicado: (2026)

NeuroFaith: Evaluating LLM Self-Explanation Faithfulness via Internal Representation Alignment
por: Bhan, Milan, et al.
Publicado: (2025)

A Comprehensive Evaluation of Neural SPARQL Query Generation from Natural Language Questions
por: Diallo, Papa Abdou Karim Karou, et al.
Publicado: (2023)

Position: Theory of Mind Benchmarks are Broken for Large Language Models
por: Riemer, Matthew, et al.
Publicado: (2024)

LLM-VA: Resolving the Jailbreak-Overrefusal Trade-off via Vector Alignment
por: Zhang, Haonan, et al.
Publicado: (2026)

Optimizing the Trade-off Between Throughput and PAoI Outage Exponents
por: Yeh, Tai-Chun, et al.
Publicado: (2025)

WorkArena++: Towards Compositional Planning and Reasoning-based Common Knowledge Work Tasks
por: Boisvert, Léo, et al.
Publicado: (2024)

Neural Coherence : Find higher performance to out-of-distribution tasks from few samples
por: Guiroy, Simon, et al.
Publicado: (2025)

Effect of Document Packing on the Latent Multi-Hop Reasoning Capabilities of Large Language Models
por: Prato, Gabriele, et al.
Publicado: (2025)

The Expressive Limits of Diagonal SSMs for State-Tracking
por: Shakerinava, Mehran, et al.
Publicado: (2026)

Intelligent Switching for Reset-Free RL
por: Patil, Darshan, et al.
Publicado: (2024)

Context-Aware Assistant Selection for Improved Inference Acceleration with Large Language Models
por: Huang, Jerry, et al.
Publicado: (2024)

Interpretability Needs a New Paradigm
por: Madsen, Andreas, et al.
Publicado: (2024)

Why Don't Prompt-Based Fairness Metrics Correlate?
por: Zayed, Abdelrahman, et al.
Publicado: (2024)

Should We Attend More or Less? Modulating Attention for Fairness
por: Zayed, Abdelrahman, et al.
Publicado: (2023)