影响力指数

64.12/100

前 3.7%

全站排名 #2,404

发表论文10 篇

平均评分5.7

年均产出3.3 篇/年

Jingxuan He

Postdoc@University of California, Berkeley·美国·OpenReview

CyberGym: Evaluating AI Agents' Real-World Cybersecurity Capabilities at Scale

Usefulness-driven Learning of Formal Mathematics

ICLR 2026Rejected

VERINA: Benchmarking Verifiable Code Generation

ICLR 2026Poster

Black-Box Adversarial Attacks on LLM-Based Code Completion

ICML 2025Poster

Mind the Gap: A Practical Attack on GGUF Quantization

ICML 2025Poster

Black-Box Adversarial Attacks on LLM-Based Code Completion

ICLR 2025Rejected

BaxBench: Can LLMs Generate Correct and Secure Backends?

ICML 2025Spotlight

合作者 (20)

博士导师7 篇

博后导师3 篇

Timothe Kasriel

Kazuki Egashira