4 papers
Understanding Diversity Collapse in RLVR via the Lens of Overtraining
Suqin Yuan, Jinkun Chen, Jiyang Zheng +6
Reinforcement learning with verifiable rewards (RLVR) has become a key approach for enhancing the reasoning abilities of large language models. However, RLVR often suffers from \em…
"I May Not Have Articulated Myself Clearly": Diagnosing Dynamic Instability in LLM Reasoning at Inference Time
Jinkun Chen, Fengxiang Cheng, Sijia Han +1
Reasoning failures in large language models (LLMs) are typically measured only at the end of a generation, yet many failures manifest as a process-level breakdown: the model "loses…
Static Sandboxes Are Inadequate: Modeling Societal Complexity Requires Open-Ended Co-Evolution in LLM-Based Multi-Agent Simulations
Jinkun Chen, Sher Badshah, Xuemin Yu +1
What if artificial agents could not just communicate, but also evolve, adapt, and reshape their worlds in ways we cannot fully predict? With llm now powering multi-agent systems an…
Addressing Correlated Latent Exogenous Variables in Debiased Recommender Systems
Shuqiang Zhang, Yuchao Zhang, Jinkun Chen +1
Recommendation systems (RS) aim to provide personalized content, but they face a challenge in unbiased learning due to selection bias, where users only interact with items they pre…