11 papers
What Should a Self-Teacher See? Privileged Context Design for On-Policy Self-Distillation
Kanghui Tian, Siyuan Liu, Tianxiang Jiang +8
More privileged information does not always make a better teacher. We study this tension in on-policy self-distillation (OPSD), where a self-teacher scores the student's own rollou…
CORTEX: High-Quality Cross-Domain Organization of Web-Scale Corpora through Ontological Corpus Graph
Chengtao Gan, Xiaoke Guo, Yushan Zhu +5
The continuous evolution of large language models drives escalating demands on data scale and quality, and as different training stages impose increasingly tailored data requiremen…
Scaling LLM Knowledge Boundaries via Distribution-Optimized Synthesis
Songze Li, Yarong Lan, Zhongpu Bo +16
Knowledge injection via synthetic data is crucial for enhancing Large Language Models (LLMs). However, current synthesis methods simply stop at preset token counts or fixed data ra…
OCELOT: Inference-Leakage Budgets for Privacy-Preserving LLM Agents
Jin Xie, Songze Li
Large language model (LLM) agents increasingly act on a user's behalf -- reading personal files, calling tools, transacting with external services -- possibly leaking personally id…
InternVideo3: Agentify Foundation Models with Multimodal Contextual Reasoning
Ziang Yan, Sheng Xia, Jiashuo Yu +10
Recent progress in foundation models has shifted toward agentic behavior involving multi-step reasoning and tool use. However, open-source efforts largely focus on text-dominant se…
What's Missing in Screen-to-Action? Towards a UI-in-the-Loop Paradigm for Multimodal GUI Reasoning
Songze Li, Xiaoke Guo, Tianqi Liu +5
Existing Graphical User Interface (GUI) reasoning tasks remain challenging, particularly in UI understanding. Current methods typically rely on direct screen-based decision-making,…
ASTRA: Adaptive Semantic Tree Reasoning Architecture for Complex Table Question Answering
Xiaoke Guo, Songze Li, Zhiqiang Liu +4
Table serialization remains a critical bottleneck for Large Language Models (LLMs) in complex table question answering, hindered by challenges such as structural neglect, represent…
Temp-R1: A Unified Autonomous Agent for Complex Temporal KGQA via Reverse Curriculum Reinforcement Learning
Zhaoyan Gong, Zhiqiang Liu, Songze Li +7
Temporal Knowledge Graph Question Answering (TKGQA) is inherently challenging, as it requires sophisticated reasoning over dynamic facts with multi-hop dependencies and complex tem…
CoG: Controllable Graph Reasoning via Relational Blueprints and Failure-Aware Refinement over Knowledge Graphs
Yuanxiang Liu, Songze Li, Xiaoke Guo +4
Large Language Models (LLMs) have demonstrated remarkable reasoning capabilities but often grapple with reliability challenges like hallucinations. While Knowledge Graphs (KGs) off…
Last Layer Logits to Logic: Empowering LLMs with Logic-Consistent Structured Knowledge Reasoning
Songze Li, Zhiqiang Liu, Zhaoyan Gong +6
Large Language Models (LLMs) achieve excellent performance in natural language reasoning tasks through pre-training on vast unstructured text, enabling them to understand the logic…
Enrich-on-Graph: Query-Graph Alignment for Complex Reasoning with LLM Enriching
Songze Li, Zhiqiang Liu, Zhengke Gui +2
Large Language Models (LLMs) exhibit strong reasoning capabilities in complex tasks. However, they still struggle with hallucinations and factual errors in knowledge-intensive scen…