11 papers
From Rollouts to Recipes: Self-Contained Post-Training for LLMs
Yifei Li, Lingling Zhang, Muye Huang +3
Post-training large language models usually applies a single training recipe to all samples, even though the model's own rollouts reveal different sample-level learning states. We…
GLaQ: Grounding Latent Queries in Visual Evidence for Multimodal Reasoning
Zesheng Yang, Lingling Zhang, Xinyu Zhang +4
Chain-of-thought reasoning has substantially improved the problem-solving capabilities of multimodal large language models. Fine-grained visual evidence, however, remains difficult…
QUMem: Personalized Memory for Query-Conditioned User-State Inference in LLM Agents
Heng Wang, Yifei Li, Lingling Zhang +4
Large language model (LLM) agents increasingly use external memory systems to support personalization by drawing on long and evolving interaction histories, in which user preferenc…
Self-Specialized Teachers for Domain Post-Training
Yifei Li, Rongman Xu, Lingling Zhang +5
Target-only post-training can improve performance in a specialized domain while degrading behaviors that a general-purpose base model acquired before adaptation. We study this prob…
Beyond Retrieval: Query-Conditioned Reuse of Long-Horizon Agent Trajectories
Yifei Li, Heng Wang, Lingling Zhang +5
Retrieval can identify a past trajectory that may matter, yet it does not specify how an acting agent should use that trajectory after users, entities, constraints, or environment…
Beyond Individual Intelligence: Surveying Collaboration, Failure Attribution, and Self-Evolution in LLM-based Multi-Agent Systems
Shihao Qi, Jie Ma, Rui Xing +15
LLM-based autonomous agents have demonstrated strong capabilities in reasoning, planning, and tool use, yet remain limited when tasks require sustained coordination across roles, t…
GeoChallenge: A Multi-Answer Multiple-Choice Benchmark for Geometric Reasoning with Diagrams
Yushun Zhang, Weiping Fu, Zesheng Yang +6
Evaluating the symbolic reasoning of large language models (LLMs) calls for geometry benchmarks that require multi-step proofs grounded in both text and diagrams. However, existing…
LogicGraph : Benchmarking Multi-Path Logical Reasoning via Neuro-Symbolic Generation and Verification
Yanrui Wu, Lingling Zhang, Xinyu Zhang +5
Evaluations of large language models (LLMs) primarily emphasize convergent logical reasoning, where success is defined by producing a single correct proof. However, many real-world…
Locomo-Plus: Beyond-Factual Cognitive Memory Evaluation Framework for LLM Agents
Yifei Li, Weidong Guo, Lingling Zhang +6
Long-term conversational memory is a core capability for LLM-based dialogue systems, yet existing benchmarks and evaluation protocols primarily focus on surface-level factual recal…
$\textbf{AGT$^{AO}$}$: Robust and Stabilized LLM Unlearning via Adversarial Gating Training with Adaptive Orthogonality
Pengyu Li, Lingling Zhang, Zhitao Gao +5
While Large Language Models (LLMs) have achieved remarkable capabilities, they unintentionally memorize sensitive data, posing critical privacy and security risks. Machine unlearni…
GeoLaux: A Benchmark for Evaluating MLLMs' Geometry Performance on Long-Step Problems Requiring Auxiliary Lines
Yumeng Fu, Jiayin Zhu, Lingling Zhang +5
Geometry problem solving (GPS) poses significant challenges for Multimodal Large Language Models (MLLMs) in diagram comprehension, knowledge application, long-step reasoning, and a…