Peipei Liu

Zhongguancun Laboratory

Papers on Lattice

Total citations

Topics

Publication activitypapers/week, last 8 weeks

Research focus

Tool Use & Agents (2)Code Generation & Program Synthesis (1)RLHF & Preference Learning (1)Reasoning & Chain-of-Thought (1)

Frequent co-authors

Congying Liu (2)Jian Sun (1)Mingzhe Xing (1)Yicheng Zeng (1)

Papers (3)

Jun 15, 2026

Zhongguancun LaboratoryJun 15, 2026·also Tsinghua AI, BIT

Binary Decompilation LLM with Feedback-Driven Multi-Turn Refinement

Iterative refinement with feedback-driven learning allows AutoDecompiler to significantly enhance the accuracy of binary decompilation, outperforming traditional single-turn models.

Peipei Liu, Jian Sun, Mingzhe Xing +5

Code Generation & Program Synthesis RLHF & Preference Learning

Apr 7, 2026

Yanxu Mao +5Apr 7, 2026·also Zhongguancun Laboratory

Stop Fixating on Prompts: Reasoning Hijacking and Constraint Tightening for Red-Teaming LLM Agents

LLM agents can be reliably jailbroken without modifying user prompts, revealing a critical vulnerability in their reasoning and memory mechanisms.

Yanxu Mao, Peipei Liu, Tiehan Cui +3

Reasoning & Chain-of-Thought Red-Teaming & Adversarial Robustness Tool Use & Agents

Mar 2, 2026

Congying Liu +5Mar 2, 2026·also Communication University of China, JHU, Zhongguancun Laboratory

ProtRLSearch: A Multi-Round Multimodal Protein Search Agent with Large Language Models Trained via Reinforcement Learning

Protein search agents can now leverage multi-dimensional reward-based RL and multimodal inputs (protein sequence + text) to produce higher quality reports for complex protein analysis tasks.

Congying Liu, Taihao Li, Ming Huang +3

Multimodal Models Scientific Discovery & Drug Design Tool Use & Agents