13 papers
UniEvo-VL: An On-policy Self-Distillation Training Recipe for Multimodal Model Self-improvement
Fang Wu, Da Xing, Yanjie Huang +16
Modern multimodal models bring generation and understanding into a single unified system, which enables them to provide and learn from their own feedback. Motivated by this unified…
MetaRSI / RSI2: A Meta-Recursive Self-Improving System for Recursive Self-Improving Systems Themselves
Zihan Tan, Leixin Sun, Zitong Shi +28
Recursive self-improvement (RSI) lets a system improve the model-building machinery from its own failures, so every later model inherits the gain. Yet RSI has been validated almost…
Emotion2Skill: Model-Internal Emotion Signals for Adaptive Skill Selection and Evolution
Bohan Lin, Hejia Geng, Xinyi Xie +5
Skill-based LLM agents select reusable procedures from external libraries to solve complex tasks, yet their routing decisions rely primarily on task descriptions, verbal reflection…
LatticeMind: A Conflict-Aware Memory Primitive for Multi-Agent Systems
Heng Zhou, Lian Zhang, Yutao Fan +5
Multi-agent LLM systems often fail not for lack of candidate answers, but because they have no persistent mechanism for deciding which incompatible claim should currently be truste…
SKT: Skill-Use Training at Scale via Verified Synthetic Data Generation
Zelin Tan, Yiqun Zhang, Hao Li +11
Agent skills have become an important mechanism for equipping language-model agents with reusable procedural knowledge. However, providing skills alone does not guarantee that curr…
Parthenon Law: A Self-Evolving Legal-Agent Framework
Hejia Geng, Leo Liu
As agents grow more capable, legal-domain LLM agents promise to turn document-heavy matters into reviewable work products -- yet reliable deployment faces three obstacles: no large…
SUDP: Secret-Use Delegation Protocol for Agentic Systems
Xiaohang Yu, Hejia Geng, Xinmeng Zeng +1
Agentic systems now act with user secrets for APIs, messaging platforms, and cloud services. Today's agent runtimes typically implement authorization by exposure: enabling action m…
Reading Seeing: Diagnosing and Closing the Typography Gap in Vision-Language Models
Heng Zhou, Ao Yu, Li Kang +5
Vision-Language Models achieve near-perfect accuracy at reading text in images, yet prove largely typography-blind: capable of recognizing what text says, but not how it looks. We…
PAPO: Stabilizing Rubric Integration Training via Decoupled Advantage Normalization
Zelin Tan, Zhouliang Yu, Bohan Lin +9
We propose Process-Aware Policy Optimization (PAPO), a method that integrates process-level evaluation into Group Relative Policy Optimization (GRPO) through decoupled advantage no…
LiveSearchBench: An Automatically Constructed Benchmark for Retrieval and Reasoning over Dynamic Knowledge
Heng Zhou, Ao Yu, Yuchen Fan +10
Evaluating large language models (LLMs) on question answering often relies on static benchmarks that reward memorization and understate the role of retrieval, failing to capture th…
Scaling Behaviors of LLM Reinforcement Learning Post-Training: An Empirical Study in Mathematical Reasoning
Zelin Tan, Hejia Geng, Xiaohang Yu +14
While scaling laws for large language models (LLMs) during pre-training have been extensively studied, their behavior under reinforcement learning (RL) post-training remains largel…
Beyond Magic Words: Sharpness-Aware Prompt Evolving for Robust Large Language Models with TARE
Guancheng Wan, Lucheng Fu, Haoxin Liu +10
The performance of Large Language Models (LLMs) hinges on carefully engineered prompts. However, prevailing prompt optimization methods, ranging from heuristic edits and reinforcem…
Diagnose, Localize, Align: A Full-Stack Framework for Reliable LLM Multi-Agent Systems under Instruction Conflicts
Guancheng Wan, Leixin Sun, Longxu Dou +10
Large Language Model (LLM)-powered multi-agent systems (MAS) have rapidly advanced collaborative reasoning, tool use, and role-specialized coordination in complex tasks. However, r…