16 papers
TSRL: Trigger-based Supervised Reinforcement Learning for Continual Agent Training
Yuqian Fu, Kaiwen Jiang, Zijun Song +3
Large language model (LLM) agents excel at complex, multi-step tasks but suffer from catastrophic forgetting during continual learning--acquiring new skills often degrades previous…
3SPO: State-Score-Supervised Policy Optimization for LLM Agents
Yu Han, Kailing Li, Yang Jiao +4
Training large language models (LLMs) as autonomous agents via reinforcement learning (RL) has enabled frontier models to achieve superhuman performance in long-horizon tasks. Howe…
When RL Fails after SFT: Rejuvenating Model Plasticity for Robust SFT-to-RL Handoff
Runze Liu, Jiashun Liu, Xu Wan +2
Supervised Fine-Tuning (SFT) followed by Reinforcement Learning (RL) has become a standard pipeline for Large Language Model (LLM) post-training. SFT is expected to provide a usefu…
Are Full Rollouts Necessary for On-Policy Distillation?
Yaocheng Zhang, Jiajun Chai, Yuqian Fu +7
On-policy distillation (OPD) provides dense teacher feedback along student-generated rollouts rather than fixed teacher traces and has emerged as a promising post-training paradigm…
A Shared Valence Axis Across Modern LLMs and Human EEG: The Saturation Regularity
Yousef A. Radwan, Xuhui Liu, Kilichbek Haydarov +2
Large language models (LLMs) have emerged as powerful representation learners whose internal features increasingly align with human cognition. We study whether modern LLMs can serv…
Bridging the 2D-3D Gap: A Hierarchical Semantic-Geometric Map for Vision Language Navigation
Kailing Li, Tianwen Qian, Lijin Yang +4
Vision-Language Navigation (VLN) enables embodied agents to reach target locations in unseen environments by following language instructions. Despite recent progress with vision-la…
Revisiting On-Policy Distillation: Empirical Failure Modes and Simple Fixes
Yuqian Fu, Haohuan Huang, Kaiwen Jiang +4
On-policy distillation (OPD) is increasingly used in LLM post-training because it can leverage a teacher model to provide dense supervision on student rollouts. The standard implem…
Is Mathematical Problem-Solving Expertise in Large Language Models Associated with Assessment Performance?
Liang Zhang, Yu Fu, Xinyi Jin
Large Language Models (LLMs) are increasingly used in math education not only as problem solvers but also as assessors of learners' reasoning. However, it remains unclear whether s…
MOSAIC: A Universal Agent-Level Interface for Cross-Paradigm Agent Mixing and Human-AI Collaboration
Abdulhamid M. Mousa, Jinhui Pang, Rakhmonberdi Khajiev +5
Existing infrastructure cannot deploy agents from different decision-making paradigms within the same environment, making fair cross-paradigm comparison under identical conditions…
VisNec: Measuring and Leveraging Visual Necessity for Multimodal Instruction Tuning
Mingkang Dong, Hongyi Cai, Jie Li +4
The effectiveness of multimodal instruction tuning depends not only on dataset scale, but critically on whether training samples genuinely require visual reasoning. However, existi…
Perception-Consistency Multimodal Large Language Models Reasoning via Caption-Regularized Policy Optimization
Songjun Tu, Qichao Zhang, Jingbo Sun +6
While multimodal large language models excel at tasks that integrate visual perception with symbolic reasoning, their performance is often undermined by a critical vulnerability: p…
SRFT: A Single-Stage Method with Supervised and Reinforcement Fine-Tuning for Reasoning
Yuqian Fu, Tinghong Chen, Jiajun Chai +7
Large language models (LLMs) have achieved remarkable progress in reasoning tasks, yet the optimal integration of Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) remai…
DipLLM: Fine-Tuning LLM for Strategic Decision-making in Diplomacy
Kaixuan Xu, Jiajun Chai, Sicheng Li +3
Diplomacy is a complex multiplayer game that requires both cooperation and competition, posing significant challenges for AI systems. Traditional methods rely on equilibrium search…
RLAE: Reinforcement Learning-Assisted Ensemble for LLMs
Yuqian Fu, Yuanheng Zhu, Jiajun Chai +4
Ensembling large language models (LLMs) can effectively combine diverse strengths of different models, offering a promising approach to enhance performance across various tasks. Ho…
Enhancing LLM Reasoning with Iterative DPO: A Comprehensive Empirical Investigation
Songjun Tu, Jiahao Lin, Xiangyu Tian +8
Recent advancements in post-training methodologies for large language models (LLMs) have highlighted reinforcement learning (RL) as a critical component for enhancing reasoning. Ho…
CPIG: Leveraging Consistency Policy with Intention Guidance for Multi-agent Exploration
Yuqian Fu, Yuanheng Zhu, Haoran Li +3
Efficient exploration is crucial in cooperative multi-agent reinforcement learning (MARL), especially in sparse-reward settings. However, due to the reliance on the unimodal policy…