English
Related papers

Related papers: Applying Refusal-Vector Ablation to Llama 3.1 70B …

200 papers

Large language models (LLMs) are typically aligned to refuse harmful instructions through safety fine-tuning. A recent attack, termed abliteration, identifies and suppresses the single latent direction most responsible for refusal behavior,…

Computation and Language · Computer Science 2025-10-08 Harethah Abu Shairah , Hasan Abed Al Kader Hammoud , Bernard Ghanem , George Turkiyyah

AI developers often apply safety alignment procedures to prevent the misuse of their AI systems. For example, before Meta released Llama 2-Chat - a collection of instruction fine-tuned large language models - they invested heavily in safety…

Machine Learning · Computer Science 2024-05-24 Simon Lermen , Charlie Rogers-Smith , Jeffrey Ladish

Training a language model to be both helpful and harmless requires careful calibration of refusal behaviours: Models should refuse to follow malicious instructions or give harmful advice (e.g."how do I kill someone?"), but they should not…

Computation and Language · Computer Science 2025-03-05 Xinpeng Wang , Chengzhi Hu , Paul Röttger , Barbara Plank

Large language models (LLMs) are vulnerable to adversarial attacks that can elicit harmful responses. Defending against such attacks remains challenging due to the opacity of jailbreaking mechanisms and the high computational cost of…

Machine Learning · Computer Science 2025-03-21 Lei Yu , Virginie Do , Karen Hambardzumyan , Nicola Cancedda

For safety reasons, large language models (LLMs) are trained to refuse harmful user instructions, such as assisting dangerous activities. We study an open question in this work: does the desired safety refusal, typically enforced in chat…

Recently, autonomous agents built on large language models (LLMs) have experienced significant development and are being deployed in real-world applications. These agents can extend the base LLM's capabilities in multiple ways. For example,…

Cryptography and Security · Computer Science 2024-07-31 Boyang Zhang , Yicong Tan , Yun Shen , Ahmed Salem , Michael Backes , Savvas Zannettou , Yang Zhang

Open-weight LLMs can be modified at inference time with simple activation edits, which raises a practical question for safety: do common safety interventions like refusal training or metatag training survive such edits? We study model…

Open-sourced Large Language Models (LLMs) have achieved great success in various NLP tasks, however, they are still far inferior to API-based models when acting as agents. How to integrate agent ability into general LLMs becomes a crucial…

Computation and Language · Computer Science 2024-03-20 Zehui Chen , Kuikun Liu , Qiuchen Wang , Wenwei Zhang , Jiangning Liu , Dahua Lin , Kai Chen , Feng Zhao

Large Language Model (LLM)-based agents are increasingly used as autonomous subordinates that carry out tasks for users. This raises the question of whether they may also engage in deception, similar to how individuals in human…

We introduce a method to improve the zero-shot reasoning abilities of large language models on general language understanding tasks. Specifically, we build an autonomous agent to instruct the reasoning process of large language models. We…

Computation and Language · Computer Science 2024-08-15 Nicholas Crispino , Kyle Montgomery , Fankun Zeng , Dawn Song , Chenguang Wang

Large Language Model (LLM) Agents are an emerging computing paradigm that blends generative machine learning with tools such as code interpreters, web browsing, email, and more generally, external resources. These agent-based systems…

Cryptography and Security · Computer Science 2024-10-23 Xiaohan Fu , Shuheng Li , Zihan Wang , Yihao Liu , Rajesh K. Gupta , Taylor Berg-Kirkpatrick , Earlence Fernandes

We propose affine concept editing (ACE) as an approach for steering language models' behavior by intervening directly in activations. We begin with an affine decomposition of model activation vectors and show that prior methods for steering…

Machine Learning · Computer Science 2025-01-29 Thomas Marshall , Adam Scherlis , Nora Belrose

Language model (LM) agents have demonstrated significant potential for automating real-world tasks, yet they pose a diverse array of potential, severe risks in safety-critical scenarios. In this work, we identify a significant gap between…

Artificial Intelligence · Computer Science 2025-08-20 Yuzhi Tang , Tianxiao Li , Elizabeth Li , Chris J. Maddison , Honghua Dong , Yangjun Ruan

This study reveals a previously unexplored vulnerability in the safety alignment of Large Language Models (LLMs). Existing aligned LLMs predominantly respond to unsafe queries with refusals, which often begin with a fixed set of prefixes…

Cryptography and Security · Computer Science 2026-01-28 Yangyang Guo , Ziwei Xu , Si Liu , Zhiming Zheng , Mohan Kankanhalli

Recent work has shown that language models' refusal behavior is primarily encoded in a single direction in their latent space, making it vulnerable to targeted attacks. Although Latent Adversarial Training (LAT) attempts to improve…

Computation and Language · Computer Science 2025-04-29 Alexandra Abbas , Nora Petrova , Helios Ael Lyons , Natalia Perez-Campanero

Rapidly evolving cyberattacks demand incident response systems that can autonomously learn and adapt to changing threats. Prior work has extensively explored the reinforcement learning approach, which involves learning response strategies…

Cryptography and Security · Computer Science 2026-04-16 Yiran Gao , Kim Hammar , Tao Li

Safety-aligned language models often refuse cybersecurity requests whose wording resembles misuse, even when the task is authorized and defensive. This makes security evaluation ambiguous: a failed answer may reflect missing capability or…

Cryptography and Security · Computer Science 2026-05-19 Isaac David , Arthur Gervais

The deployment of Large Language Models (LLMs) as tool-using agents causes their alignment training to manifest in new ways. Recent work finds that language models can use tools in ways that contradict the interests or explicit instructions…

Machine Learning · Computer Science 2026-04-24 Kushal Agrawal , Frank Xiao , Guido Bergman , Asa Cooper Stickland

Safety-aligned language models systematically refuse harmful requests. While activation steering can modulate refusal, ablating the raw "refusal vector" calculated from contrastive harmful and harmless prompts often causes collateral damage…

Computation and Language · Computer Science 2026-01-14 Tony Cristofano

Safety-aligned language models refuse harmful requests through learned refusal behaviors encoded in their internal representations. Recent activation-based jailbreaking methods circumvent these safety mechanisms by applying orthogonal…

Machine Learning · Computer Science 2026-03-05 Geraldin Nanfack , Eugene Belilovsky , Elvis Dohmatob
‹ Prev 1 2 3 10 Next ›