Xianfeng Tang

How Far Are LLMs from Professional Poker Players? Revisiting Game-Theoretic Reasoning with Agentic Tool Use

Agent-REINFORCE: Searching Compute-Optimal Multi-LLM Collaboration Graph for Test-Time Scaling

Unlocking the Power of Multi-Agent LLM for Reasoning: From Lazy Agents to Deliberation

TRAJECT-Bench:A Trajectory-Aware Benchmark for Evaluating Agentic Tool Use

Seeing but Not Believing: Probing the Disconnect Between Visual Attention and Answer Correctness in VLMs

Inference Scaling of LLM Ensembling: Bridging Token Spaces with Token Translation

Bradley-Terry and Multi-Objective Reward Modeling Are Complementary

Beyond Chunks and Graphs: Retrieval-Augmented Generation through Triplet-Driven Thinking

A New Paradigm for Genome-wide DNA Methylation Prediction Without Methylation Input

Synthesizing High-Quality Visual Question Answering from Medical Documents with Generator-Verifier LMMs

All You Need Are Random Visual Tokens? Demystifying Token Pruning in VLLMs

Where in the World? A Vision-Language Benchmark for Probing Model Geolocation Skills Across Scales

Node-wise Filtering in Graph Neural Networks: A Mixture of Experts Approach

Adaptive Test-Time Reasoning via Reward-Guided Dual-Phase Search

Keeping an Eye on LLM Unlearning: The Hidden Risk and Remedy

AgentTTS: Large Language Model Agent for Test-time Compute-optimal Scaling Strategy in Complex Tasks

RRO: LLM Agent Optimization Through Rising Reward Trajectories

Catastrophic Failure of LLM Unlearning via Quantization

Mitigating Heterogeneous Token Overfitting in LLM Knowledge Editing

Unlocking Efficient, Scalable, and Continual Knowledge Editing with Basis-Level Representation Fine-Tuning

Automatic Task-aware Instruction Optimizer for Black-box LLMs

Divide-Verify-Refine: Aligning LLM Responses with Complex Instructions

Node-wise Filtering in Graph Neural Networks: A Mixture of Experts Approach

Robust Watermarking for Diffusion Models: A Unified Multi-Dimensional Recipe

Two-Step Offline Preference-Based Reinforcement Learning with Constrained Actions