@inproceedings{yang-etal-2026-toolcpt,
    title = "{T}ool{CPT}: Improving Tool Utilization in {LLM} Agents via Continuous Pre-training",
    author = "Yang, Yifan  and
      Lu, Jinghui  and
      Evadeng  and
      Yang, Ao  and
      Yu, Peijie  and
      YU, TingHao  and
      Zhang, Feng",
    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.776/",
    pages = "15830--15856",
    ISBN = "979-8-89176-395-1",
    abstract = "Autonomous agents powered by large language models (LLM-based agents) are capable of using off-the-shelf tools to interact with the environment, solve real-world problems, and boost work efficiency. However, current approaches to enhancing tool use for LLM-based agents primarily focus on post-training fine-tuning or test-time context extension. These methods overlook the fundamental tool knowledge acquisition during the early training phase, where models actually learn and internalize core knowledge representations, restricting model performance on out-of-distribution tool usage. To solve such a problem, we introduce enhancing \textbf{tool} knowledge for LLM-based agents during \textbf{c}ontinuous \textbf{p}re-\textbf{t}raining (\textbf{ToolCPT}). We identify and bridge a key gap in current LLM training by shifting focus from tool-calling patterns to deep internalization of core tool-knowledge representations. We begin by curating 5.1 million code artifacts from large-scale, high-quality code repositories. These artifacts are selected based on a set of criteria that defines a usable ``proxy agent tool'', thereby forming a comprehensive agent tool library. For each proxy tool, we then create a detailed playbook covering implementation specifications, core functionalities, interaction protocols with other tools, and illustrative positive and negative examples. This process yields a large-scale tool knowledge corpus comprising 18 billion tokens, which is used to continuously pre-train our model. Experiments show our playbook-enhanced corpus catalyzes deep knowledge internalization, driving the model to notable performance gains on multiple standard benchmarks."
}