Publications (23)
Unsafer in Many Turns: Benchmarking and Defending Multi-Turn Safety Risks in Tool-Using Agents
Xu Li, Simon Yu, Minzhou Pan +5
LLM-based agents are becoming increasingly capable, yet their safety lags behind. This creates a gap between what agents can do and should do. This gap widens as agents engage in m…
TextArena
Leon Guertler, Bobby Cheng, Simon Yu +3
TextArena is an open-source collection of competitive text-based games for training and evaluation of agentic behavior in Large Language Models (LLMs). It spans 57+ unique environm…
Verbalized Sampling: How to Mitigate Mode Collapse and Unlock LLM Diversity
Jiayi Zhang, Simon Yu, Derek Chong +4
The paper identifies typicality bias in preference data as a key cause of mode collapse in aligned large language models and introduces Verbalized Sampling, a training‑free prompti…
PolySkill: Learning Generalizable Skills Through Polymorphic Abstraction
Simon Yu, Gang Li, Weiyan Shi +1
Large language models (LLMs) are moving beyond static uses and are now powering agents that learn continually during their interaction with external environments. For example, agen…
Instances and Labels: Hierarchy-aware Joint Supervised Contrastive Learning for Hierarchical Multi-Label Text Classification
Simon Yu, Jie He, VÃctor Gutiérrez-Basulto +1
Hierarchical multi-label text classification (HMTC) aims at utilizing a label hierarchy in multi-label classification. Recent approaches to HMTC deal with the problem of imposing a…
Is It Good Data for Multilingual Instruction Tuning or Just Bad Multilingual Evaluation for Large Language Models?
Pinzhen Chen, Simon Yu, Zhicheng Guo +1
Multilingual large language models are designed, claimed, and expected to cater to speakers of varied languages. We hypothesise that the current practices of fine-tuning and evalua…
LiDAR Cluster First and Camera Inference Later: A New Perspective Towards Autonomous Driving
Jiyang Chen, Simon Yu, Rohan Tabish +4
Object detection in state-of-the-art Autonomous Vehicles (AV) framework relies heavily on deep neural networks. Typically, these networks perform object detection uniformly on the…
Shepherd: Enabling Programmable Meta-Agents via Reversible Agentic Execution Traces
Simon Yu, Derek Chong, Ananjan Nandi +4
As LLM agent systems take on more complex tasks, they increasingly rely on meta-agents: higher-order agents that create, operate on and manage other agents. Meta-agent operations s…
WHEN TO ACT, WHEN TO WAIT: Modeling the Intent-Action Alignment Problem in Dialogue
Yaoyao Qian, Jindan Huang, Yuanli Wang +5
Dialogue systems often fail when user utterances are semantically complete yet lack the clarity and completeness required for appropriate system action. This mismatch arises becaus…
Coding with "Enemy": Can Human Developers Detect AI Agent Sabotage?
Jingheng Ye, Huiqi Zou, Simon Yu +1
AI coding agents are increasingly embedded in real-world software development, collaborating with human developers while gaining broader access to codebases and tools. This creates…
Fine-tuning Large Language Models with Sequential Instructions
Hanxu Hu, Simon Yu, Pinzhen Chen +1
Despite the success of existing instruction-tuned models, we find that they usually struggle to respond to queries with multiple instructions. This impairs their performance in com…
SPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learning
Bo Liu, Leon Guertler, Simon Yu +9
Recent advances in reinforcement learning have shown that language models can develop sophisticated reasoning through training on tasks with verifiable rewards, but these approache…
Segmentation and Vascular Vectorization for Coronary Artery by Geometry-based Cascaded Neural Network
Xiaoyu Yang, Lijian Xu, Simon Yu +3
Segmentation of the coronary artery is an important task for the quantitative analysis of coronary computed tomography angiography (CCTA) images and is being stimulated by the fiel…
VISAT: Benchmarking Adversarial and Distribution Shift Robustness in Traffic Sign Recognition with Visual Attributes
Simon Yu, Peilin Yu, Hongbo Zheng +3
We present VISAT, a novel open dataset and benchmarking suite for evaluating model robustness in the task of traffic sign recognition with the presence of visual attributes. Built…
Neural Probabilistic Circuits: Enabling Compositional and Interpretable Predictions through Logical Reasoning
Weixin Chen, Simon Yu, Huajie Shao +2
End-to-end deep neural networks have achieved remarkable success across various domains but are often criticized for their lack of interpretability. While post hoc explanation meth…
N2C2: Nearest Neighbor Enhanced Confidence Calibration for Cross-Lingual In-Context Learning
Jie He, Simon Yu, Deyi Xiong +2
Recent advancements of in-context learning (ICL) show language models can significantly improve their performance when demonstrations are provided. However, little attention has be…
Perception Simplex: Verifiable Collision Avoidance in Autonomous Vehicles Amidst Obstacle Detection Faults
Ayoosh Bansal, Hunmin Kim, Simon Yu +4
Advances in deep learning have revolutionized cyber-physical applications, including the development of Autonomous Vehicles. However, real-world collisions involving autonomous con…
Verifiable Obstacle Detection
Ayoosh Bansal, Hunmin Kim, Simon Yu +4
Perception of obstacles remains a critical safety concern for autonomous vehicles. Real-world collisions have shown that the autonomy faults leading to fatal collisions originate f…
Overconfident and Blind to Details: Fixing Prompt Insensitivity with Abductive Preference Learning
Yijin Ni, Simon Yu, Peng Qi
Vision and language models frequently ignore semantically critical input edits, defaulting to pretraining priors. For example, models will confidently assert a five-legged dog has…
Evaluating and Safeguarding the Adversarial Robustness of Retrieval-Based In-Context Learning
Simon Yu, Jie He, Pasquale Minervini +1
With the emergence of large language models, such as LLaMA and OpenAI GPT-3, In-Context Learning (ICL) gained significant attention due to its effectiveness and efficiency. However…
Voxel2Hemodynamics: An End-to-end Deep Learning Method for Predicting Coronary Artery Hemodynamics
Ziyu Ni, Linda Wei, Lijian Xu +4
Local hemodynamic forces play an important role in determining the functional significance of coronary arterial stenosis and understanding the mechanism of coronary disease progres…
Diversify and Conquer: Diversity-Centric Data Selection with Iterative Refinement
Simon Yu, Liangyu Chen, Sara Ahmadian +1
Finetuning large language models on instruction data is crucial for enhancing pre-trained knowledge and improving instruction-following capabilities. As instruction datasets prolif…
GEM: A Gym for Agentic LLMs
Zichen Liu, Anya Sims, Keyu Duan +16
The training paradigm for large language models (LLMs) is moving from static datasets to experience-based learning, where agents acquire skills via interacting with complex environ…