Zhuosheng Zhang
Assistant Professor
Works on natural language processing, AI agents and agent safety.
Zhuosheng Zhang's lab at SJTU works on reasoning & agents and interpretability, evaluation & safety of LMs. 42 papers since 2023, 35 at top venues, mostly EMNLP and ACL. 6 PhD students and 8 master's students. 7 alumni, most often to Alibaba and Tencent. Applications close .
At a glance
- papers, 2023–2026
- 11/ year
- at T1 venues
- 83%
- PhD students
- 6
- MS students
- 8
- alumni listed
- 7
Papers by yearfrom the lab homepage, 2023–2026
- 36 papers2026 · T1 29 · collecting
- 2 papers2025 · T1 2
- 2 papers2024 · T1 2
- 2 papers2023 · T1 2
T1 venuesOtherVenues: EMNLP, ACL, CVPR, ICML, NeurIPS
Research focuspaper titles that use each topic's words
- Reasoning & agents22 papers
- agents16 papers
- agent5 papers
- Interpretability, evaluation & safety of LMs2 papers
- safety2 papers
Paper list
From the lab homepage.
202636 papers
- ACL (T1)Agent-Dice: Disentangling Knowledge Updates via Geometric Consensus for Agent Continual Learning
- AAAI (T1)An LLM-based Quantitative Framework for Evaluating High-Stealthy Backdoor Risks in OSS Supply Chains
- IEEE Transactions on Fuzzy SystemsAuditing Partial Dataset Usage in Large Language Models Via Fuzzy Membership Aggregation
- TASLP (T1)Beyond Elicitation: Provision-Based Prompt Optimization for Knowledge-Intensive Tasks
- EMNLP (T1)Causal Probing for Internal Visual Representations in Multimodal Large Language Models
- EMNLP (T1)Chain-of-Trigger: An Agentic Backdoor that Paradoxically Enhances Agentic Robustness
- WWW (T1)ColorBench: Benchmarking Mobile Agents with Graph-Structured Framework for Complex Long-Horizon Tasks
- ACL Industry TrackColorBrowserAgent: Complex Long-Horizon Browser Agent with Adaptive Knowledge Evolution
- EMNLP (T1)Communication Policy Evolution for Proactive LLM Agents
- ICLR (T1)DeepMath-103K: A Large-Scale, Challenging, Decontaminated, and Verifiable Mathematical Dataset for Advancing Reasoning
- IJCAI-ECAIDiG-Plan: Mitigating Early Commitment for Tool-Graph Planning via Diffusion Guidance
- TPAMI (T1)Discourse-Aware Language Representation
20252 papers
- Nature Communications (T1)Risks of AI Scientists: Prioritizing Safeguarding Over Autonomy
- NeurIPS (T1)Thoughts Are All Over the Place: On the Underthinking of Long Reasoning Models
20242 papers
- TMLR (T1)Multimodal Chain-of-Thought Reasoning in Language Models
- EMNLP (T1)R-Judge: Benchmarking Safety Risk Awareness for LLM Agents
20232 papers
- ICLR (T1)Automatic Chain of Thought Prompting in Large Language Models
- EMNLP (T1)Is ChatGPT a General-Purpose Natural Language Processing Task Solver?
Where alumni went7 listed, 40% to industry
Industry 2Academia 3
Most often to
- Alibaba
- Tencent
Also to
- NUS
- Shanghai University
Apply to SJTUFall 2027
Computer Science PhD (international applicants)
- Round
- Self-funded
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- GBCambridge
Bill Byrne
- Multimodal & grounded language
- USCMU
Alexander Waibel
- Speech recognition & spoken language
- GBCambridge
Phil Woodland
- Speech recognition & spoken language
- GBCambridge
Simone Teufel
2 T1 - GBEdinburgh
Sharon Goldwater
- Multilingual, translation & linguistics
- USUMD
Naomi Feldman
- Multimodal & grounded language
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
SJTUcloses