影响力指数

72.74/100

前 2.2%

全站排名 #1,421

发表论文14 篇

平均评分5.8

年均产出4.7 篇/年

Matt Fredrikson

Researcher@Gray Swan AI·美国·OpenReview

研究方向

AI Safety · AI Security · Robustness · Adversarial Attacks · Verification · Security & Privacy · Programming Languages

LipNeXt: Scaling up Lipschitz-based Certified Robustness to Billion-parameter Models

ICLR 2026Poster

Comparing AI Agents to Cybersecurity Professionals in Real-World Penetration Testing

ICLR 2026Poster

A Mixture of Linear Corrections Generates Secure Code

ICLR 2026Rejected

Aligned LLMs Are Not Aligned Browser Agents

ICLR 2025Poster

Safety Pretraining: Toward the Next Generation of Safe AI

NeurIPS 2025Poster

AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents

ICLR 2025Poster

LLM-based Multi-Agents System Attack via Continuous Optimization with Discrete Efficient Search

COLM 2025Poster

Transferable Adversarial Attack on Vision-enabled Large Language Models

ICLR 2025Withdrawn

合作者 (20)

PhD Advisee7 篇

PhD Advisee5 篇

PhD Advisee4 篇

PhD Advisee4 篇