影响力指数

14.52/100

超过 44.3%

全站排名 #35,835

发表论文6 篇

平均评分4.5

年均产出2.0 篇/年

Zhe Li

Researcher@Singapore Management University·新加坡·OpenReview

Where Did It Go Wrong? Attributing Undesirable LLM Behaviors via Representation Gradient Tracing

ICLR 2026Poster

AutoBackdoor: Automating Backdoor Attacks via LLM Agents

ICLR 2026Rejected

Adversarial Suffixes May Be Features Too!

ICLR 2025Withdrawn

Do Influence Functions Work on Large Language Models?

ICLR 2025Withdrawn

合作者 (13)

博士导师4 篇