Nicholas Meade

The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents

arXiv 2026

SafeArena: Evaluating the Safety of Autonomous Web Agents

arXiv 2025

AgentRewardBench: Evaluating Automatic Evaluations of Web Agent Trajectories

arXiv 2025

DeepSeek-R1 Thoughtology: Let's think about LLM Reasoning

arXiv 2025

Exploiting Instruction-Following Retrievers for Malicious Information Retrieval

arXiv 2025

Universal Adversarial Triggers Are Not Universal

arXiv 2024

Evaluating Correctness and Faithfulness of Instruction-Following Models for Question Answering

arXiv 2023

An Empirical Survey of the Effectiveness of Debiasing Techniques for Pre-trained Language Models

ACL 2022 5

Evaluating the Faithfulness of Importance Measures in NLP by Recursively Masking Allegedly Important Tokens and Retraining

arXiv 2021

Siva Reddy

Arkil Patel

Xing Han Lù

Karolina Stańczak

Parishad BehnamGhader

Vaibhav Adlakha

Alejandra Zambrano

Amirhossein Kazemnejad

Dongchan Shin

researcher

Ada Defne Tur