Zeming Wei

@weizeming25

Ph.D. Candidate @PKU1898, Research Intern @BytedanceTalk Seed, Ex-Visiting Student @UCBerkeley. I focus on developing Trustworthy AI/ML.

Beijing

weizeming.github.io

Born 2002

Replies Reposts Media

Pinned
Zeming Wei
@weizeming25
Aug 21
🚨 Our paper SkillSafetyBench is accepted to #EMNLP2026 Main! Agent skills are becoming the plugin ecosystem of LLM agents. We ask a simple question: what if the attacker is not the user, but the skill itself? 📄 arxiv.org/abs/2605.12015
2
Zeming Wei
@weizeming25
Sep 5, 2025
🚨False Sense of Security: Our new paper identifies a critical limitation in representation probing-based malicious input detection—purported "high detection accuracy" may confer a false sense of security: arxiv.org/pdf/2509.03888
1
Zeming Wei
@weizeming25
Jul 22, 2024
We will present our #ICML2024 poster tomorrow at Hall C 4-9 #2411 🥳 Welcome to have a look Paper: On the Duality Between Sharpness-Aware Minimization and Adversarial Training PDF: arxiv.org/pdf/2402.15152 Code: github.com/weizeming/SAM_…
1
Zeming Wei
@weizeming25
Apr 4, 2024
Glad to share that I just reached 100 citations according to Google Scholar. Thanks all coauthors!
scholar.google.com
Zeming Wei - Cited by 1,556
Peking University | ByteDance Seed - Trustworthy AI - Safety Alignment - Robust Generalization
2
Zeming Wei
@weizeming25
Dec 31, 2023
🧵(1/8) I've completed my studies at UC Berkeley and resumed my undergraduate studies at Peking University 😆 ✨ Some highlights from my visit to Berkeley include:
Zeming Wei
@weizeming25
Aug 28, 2023
Update: I’ve just started a visiting student program at UC Berkeley for 4 months. Please DM me if you’re also here. Looking forward to meeting old and new friends!
3