7 papers · 1 filter
SafeAgent: A Runtime Protection Architecture for Agentic Systems
Hailin Liu, Eugene Ilyushin, Jie Ni +1
Large language model (LLM) agents are vulnerable to prompt-injection attacks that propagate through multi-step workflows, tool interactions, and persistent context, making input-ou…
DeepReviewer 2.0: A Traceable Agentic System for Auditable Scientific Peer Review
Yixuan Weng, Minjun Zhu, Qiujie Xie +7
Automated peer review is often framed as generating fluent critique, yet reviewers and area chairs need judgments they can \emph{audit}: where a concern applies, what evidence supp…
AutoFigure: Generating and Refining Publication-Ready Scientific Illustrations
Minjun Zhu, Zhen Lin, Yixuan Weng +6
High-quality scientific illustrations are crucial for effectively communicating complex scientific and technical concepts, yet their manual creation remains a well-recognized bottl…
Abduct, Act, Predict: Scaffolding Causal Inference for Automated Failure Attribution in Multi-Agent Systems
Alva West, Yixuan Weng, Minjun Zhu +3
Failure attribution in multi-agent systems -- pinpointing the exact step where a decisive error occurs -- is a critical yet unsolved challenge. Current methods treat this as a patt…
How Far Are AI Scientists from Changing the World?
Qiujie Xie, Yixuan Weng, Minjun Zhu +9
The emergence of large language models (LLMs) is propelling automated scientific discovery to the next level, with LLM-based Artificial Intelligence (AI) Scientist systems now taki…
Constrain Alignment with Sparse Autoencoders
Qingyu Yin, Chak Tou Leong, Minjun Zhu +7
The alignment of large language models (LLMs) with human preferences remains a key challenge. While post-training techniques like Reinforcement Learning from Human Feedback (RLHF)…