11 citations · 20 across the 10 of their papers we have counts for
10 papers
Evaluating AI Models' Capability to Automate Voice Phishing Attacks
Fred Heiding, Claudio Mayrink Verdun, Simon Lermen +5
Voice phishing (vishing) attacks have traditionally been limited by the need for human operators. The rapid emergence of high-quality AI voice synthesis and large language models (…
Large-scale online deanonymization with LLMs
Simon Lermen, Daniel Paleka, Joshua Swanson +3
We show that large language models can be used to perform at-scale deanonymization. With full Internet access, our agent can re-identify Hacker News users and Anthropic Interviewer…
Can AI Models be Jailbroken to Phish Elderly Victims? An End-to-End Evaluation
Fred Heiding, Simon Lermen
We present an end-to-end demonstration of how attackers can exploit AI safety failures to harm vulnerable populations: from jailbreaking LLMs to generate phishing content, to deplo…
Deceptive Automated Interpretability: Language Models Coordinating to Fool Oversight Systems
Simon Lermen, Mateusz Dziemian, Natalia Pérez-Campanero Antolín
We demonstrate how AI agents can coordinate to deceive oversight systems using automated interpretability of neural networks. Using sparse autoencoders (SAEs) as our experimental f…
Evaluating Large Language Models' Capability to Launch Fully Automated Spear Phishing Campaigns: Validated on Human Subjects
Fred Heiding, Simon Lermen, Andrew Kao +2
In this paper, we evaluate the capability of large language models to conduct personalized phishing attacks and compare their performance with human experts and AI models from last…
Applying Refusal-Vector Ablation to Llama 3.1 70B Agents
Simon Lermen, Mateusz Dziemian, Govind Pimpale
Recently, language models like Llama 3.1 Instruct have become increasingly capable of agentic behavior, enabling them to perform tasks requiring short-term planning and tool use. I…
Exploring the Robustness of Model-Graded Evaluations and Automated Interpretability
Simon Lermen, Ondřej Kvapil
There has been increasing interest in evaluations of language models for a variety of risks and characteristics. Evaluations relying on natural language understanding for grading c…
BadLlama: cheaply removing safety fine-tuning from Llama 2-Chat 13B
Pranav Gade, Simon Lermen, Charlie Rogers-Smith +1
Llama 2-Chat is a collection of large language models that Meta developed and released to the public. While Meta fine-tuned Llama 2-Chat to refuse to output harmful content, we hyp…
LoRA Fine-tuning Efficiently Undoes Safety Training in Llama 2-Chat 70B
Simon Lermen, Charlie Rogers-Smith, Jeffrey Ladish
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-t…
Evaluating Shutdown Avoidance of Language Models in Textual Scenarios
Teun van der Weij, Simon Lermen, Leon lang
Recently, there has been an increase in interest in evaluating large language models for emergent and dangerous capabilities. Importantly, agents could reason that in some scenario…