5 papers
AdaKP: Online Adaptive Knowledge-Point Selection for Reasoning-Oriented Reinforcement Learning
Zibin Meng, Zhenyu Zhao, Chunqiang Run
Reinforcement learning with verifiable rewards is a powerful paradigm for eliciting reasoning in large language models, yet it suffers from severe reward sparsity on competition-le…
CRAFT: Counterfactual Credit Assignment from Free Sibling Rollouts for Self-Distilled Agentic Reinforcement Learning
Zibin Meng, Kani Chen
Self-distilled agentic reinforcement learning augments trajectory-level reward with a token-level distillation loss, using as its teacher the same policy conditioned on privileged…
Depth-Entropy Guided Sampling for Training-Free LLM Reasoning
Zibin Meng, Peng Xie, Kani Chen
Reinforcement learning (RL) has become the dominant paradigm for improving the reasoning capabilities of large language models, but it requires expensive training, curated data, an…
PsyAgent: Constructing Human-like Agents Based on Psychological Modeling and Contextual Interaction
Zibin Meng, Kani Chen
Human-like agents must express stable dispositions while adapting to roles, relationships, and norms. We present PsyAgent, a schema-first framework that operationalizes the trait-c…
SocialGPT: Prompting LLMs for Social Relation Reasoning via Greedy Segment Optimization
Wanhua Li, Zibin Meng, Jiawei Zhou +3
Social relation reasoning aims to identify relation categories such as friends, spouses, and colleagues from images. While current methods adopt the paradigm of training a dedicate…