@inproceedings{tang-etal-2026-implicit,
    title = "From Implicit Graph Encoding to Explicit Evidence: A Training-Free {LLM} Framework for Temporal Knowledge Graph Reasoning",
    author = "Tang, Guo  and
      Cheng, Ke  and
      Fan, Huiming  and
      Chang, Heng  and
      Zheng, Wenxiang  and
      Ou, Xianhao  and
      Xiang, Junjia  and
      Liu, Ming  and
      Zhou, Yujun  and
      Lanyu, Li  and
      Qin, Bing",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.266/",
    pages = "5383--5405",
    ISBN = "979-8-89176-395-1",
    abstract = "Temporal Knowledge Graph (TKG) forecasting faces significant challenges due to distribution shifts and poor inductive generalization in parametric models. While Large Language Models (LLMs) offer potent semantic reasoning, existing LLM-based approaches struggle with implicit modality alignment and suboptimal graph linearization, failing to capture complex topologies without retraining. To bridge this gap, we propose ExE-LLM, a training-free, test-time adaptive framework that reframes TKG prediction as explicit evidence-driven reasoning. Our core philosophy is to decouple topological calculation from semantic reasoning: a heuristic module translates latent graph signals into natural language evidence, enabling the LLM to perform multi-source judgment. ExE-LLM incorporates a task-aware scheduler for test-time adaptation, a heuristic synthesizer for explicit modality alignment, and a self-diagnosis module for iterative optimization. Extensive experiments on four benchmarks demonstrate that ExE-LLM achieves SOTA performance in inductive settings, significantly outperforming fully trained graph neural networks without updating LLM parameters. The source code is available at https://github.com/JENLISA4EVER/ExE-LLM."
}