@inproceedings{duan-etal-2026-memtr,
    title = "{M}em{TR}: Enhancing Tool-Calling Reliability via Uncertainty-Triggered {FFN}-Space Retracing",
    author = "Duan, Hongtao  and
      Jiang, Lu  and
      Zhang, Minying  and
      Zhu, Xiaobing  and
      Bu, Tianpeng  and
      Jiang, Hao  and
      Wei, Xinyu  and
      hu, Lulu",
    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.973/",
    pages = "19476--19493",
    ISBN = "979-8-89176-395-1",
    abstract = "Tool calling requires Large Language Models (LLMs) to generate structured decisions including tool names and schema-constrained arguments, where small decoding mistakes can cause hard failures. Existing methods either rely on costly tool-use training data or only constrain syntax, leaving tool selection and argument value errors largely unsolved. We analyze tool calling failures through a Where{--}When lens: (Where) failures correlate with persistent uncertainty in late transformer layers, (When) uncertainty concentrates on content-bearing tokens (tool names and argument values) rather than schema tokens. Based on this, and motivated by evidence that transformer Feed Forward Networks (FFNs) act as key{--}value style memories that store and retrieve factual or associative mappings, we propose Memory Space Tool Retracing (MemTR), a weight-free decoding-time method that retrieves relevant tool evidence from the tool library and mixes it into the FFN-output at the uncertain layer, treating FFNs as key{--}value memories. Through extensive experiments on various model families (Qwen, Llama, and xLAM) and benchmarks (BFCL, ACEBench, APIBank), MemTR reduces tool calling failures by 2{\%}{--}9{\%} with only 1{\%}{--}2{\%} runtime overhead, without any fine-tuning or additional tool-use training data."
}