{
  "annotation_id": "950227cf-d5c0-4d6d-938e-a58e85adaf44",
  "date_created": "2026-02-17T05:53:24.050000Z",
  "date_modified": "2026-02-17T05:53:24.050000Z",
  "file_hash": "d61033ae6343efae540d51858cd2c28918ab02a358b07f2b7574bb33028e57fd",
  "private": false,
  "record": {
    "abstract": "Tool-augmented large language models (LLMs) have powered many applications. However, they are likely to suffer from knowledge conflict. In this paper, we propose a new type of knowledge conflict -- Tool-Memory Conflict (TMC), where the internal parametric knowledge contradicts with the external tool knowledge for tool-augmented LLMs. We find that existing LLMs, though powerful, suffer from TMC, especially on STEM-related tasks. We also uncover that under different conditions, tool knowledge and parametric knowledge may be prioritized differently. We then evaluate existing conflict resolving techniques, including prompting-based and RAG-based methods. Results show that none of these approaches can effectively resolve tool-memory conflicts.",
    "arxiv_id": "2601.09760",
    "authors": [
      "Jiali Cheng",
      "Rui Pan",
      "Hadi Amiri"
    ],
    "categories": [
      "cs.SE",
      "cs.AI"
    ],
    "license": "http://creativecommons.org/licenses/by/4.0/",
    "title": "Investigating Tool-Memory Conflicts in Tool-Augmented LLMs",
    "url": "https://arxiv.org/abs/2601.09760",
    "version": "v1"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "8c60ba08-8ad2-4fe8-864f-9c3c5062f93d",
    "id": "arXiv Dataset",
    "type": "Model",
    "variant": "snapshot-2026-01-17",
    "version": "0.1.0"
  },
  "user_id": 1000002
}