Rethinking Stateful Tool Use in Multi-Turn Dialogues: Benchmarks and Challenges

• URL: http://arxiv.org/abs/2505.13328v1
• Date: Mon, 19 May 2025 16:36:13 GMT
• Title: Rethinking Stateful Tool Use in Multi-Turn Dialogues: Benchmarks and Challenges
• Authors: Hongru Wang, Wenyu Huang, Yufei Wang, Yuanhao Xi, Jianqiao Lu, Huan Zhang, Nan Hu, Zeming Liu, Jeff Z. Pan, Kam-Fai Wong,
• Abstract summary: Existing benchmarks that assess Language Models (LMs) as Language Agents (LAs) for tool use primarily focus on single-turn interactions.<n>We propose textttDialogTool, a multi-turn dialogue dataset with stateful tool interactions considering the whole life cycle of tool use.
• Score: 30.68589269821412
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Existing benchmarks that assess Language Models (LMs) as Language Agents (LAs) for tool use primarily focus on stateless, single-turn interactions or partial evaluations, such as tool selection in a single turn, overlooking the inherent stateful nature of interactions in multi-turn applications. To fulfill this gap, we propose \texttt{DialogTool}, a multi-turn dialogue dataset with stateful tool interactions considering the whole life cycle of tool use, across six key tasks in three stages: 1) \textit{tool creation}; 2) \textit{tool utilization}: tool awareness, tool selection, tool execution; and 3) \textit{role-consistent response}: response generation and role play. Furthermore, we build \texttt{VirtualMobile} -- an embodied virtual mobile evaluation environment to simulate API calls and assess the robustness of the created APIs\footnote{We will use tools and APIs alternatively, there are no significant differences between them in this paper.}. Taking advantage of these artifacts, we conduct comprehensive evaluation on 13 distinct open- and closed-source LLMs and provide detailed analysis at each stage, revealing that the existing state-of-the-art LLMs still cannot perform well to use tools over long horizons.

Related papers

• ToolHaystack: Stress-Testing Tool-Augmented Language Models in Realistic Long-Term Interactions [9.825432101000358]
  We introduce ToolHaystack, a benchmark for testing the tool use capabilities in long-term interactions.<n>Each test instance includes multiple tasks execution contexts and realistic noise within a continuous conversation.<n>We find that while current models perform well in standard multi-turn settings, they often significantly struggle in ToolHaystack.
  arXiv  Detail & Related papers  (2025-05-29T17:10:12Z)
• Advancing and Benchmarking Personalized Tool Invocation for LLMs [66.39214525683425]
  We introduce the concept of Personalized Tool Invocation and define two key tasks: Tool Preference and Profile-dependent Query.<n>To tackle these challenges, we propose PTool, a data synthesis framework designed for personalized tool invocation.<n>We construct textbfPTBench, the first benchmark for evaluating personalized tool invocation.
  arXiv  Detail & Related papers  (2025-05-07T02:25:20Z)
• ACEBench: Who Wins the Match Point in Tool Usage? [68.54159348899891]
  ACEBench is a comprehensive benchmark for assessing tool usage in Large Language Models (LLMs)<n>It categorizes data into three primary types based on evaluation methodology: Normal, Special, and Agent.<n>It provides a more granular examination of error causes across different data types.
  arXiv  Detail & Related papers  (2025-01-22T12:59:08Z)
• ToolSandbox: A Stateful, Conversational, Interactive Evaluation Benchmark for LLM Tool Use Capabilities [30.030101957186595]
  ToolSandbox is an evaluation framework for large language models (LLMs)<n>ToolSandbox includes stateful tool execution, implicit state dependencies between tools, a built-in user simulator supporting on-policy conversational evaluation.<n>We show that open source and proprietary models have a significant performance gap, and complex tasks like State Dependency, Canonicalization and Insufficient Information defined in ToolSandbox are challenging even the most capable SOTA LLMs.
  arXiv  Detail & Related papers  (2024-08-08T05:45:42Z)
• Re-Invoke: Tool Invocation Rewriting for Zero-Shot Tool Retrieval [47.81307125613145]
  Re-Invoke is an unsupervised tool retrieval method designed to scale effectively to large toolsets without training. We employ a novel multi-view similarity ranking strategy based on intents to pinpoint the most relevant tools for each query. Our evaluation demonstrates that Re-Invoke significantly outperforms state-of-the-art alternatives in both single-tool and multi-tool scenarios.
  arXiv  Detail & Related papers  (2024-08-03T22:49:27Z)
• Towards Completeness-Oriented Tool Retrieval for Large Language Models [60.733557487886635]
  Real-world systems often incorporate a wide array of tools, making it impractical to input all tools into Large Language Models. Existing tool retrieval methods primarily focus on semantic matching between user queries and tool descriptions. We propose a novel modelagnostic COllaborative Learning-based Tool Retrieval approach, COLT, which captures not only the semantic similarities between user queries and tool descriptions but also takes into account the collaborative information of tools.
  arXiv  Detail & Related papers  (2024-05-25T06:41:23Z)
• Planning, Creation, Usage: Benchmarking LLMs for Comprehensive Tool Utilization in Real-World Complex Scenarios [93.68764280953624]
  UltraTool is a novel benchmark designed to improve and evaluate Large Language Models' ability in tool utilization. It emphasizes real-world complexities, demanding accurate, multi-step planning for effective problem-solving. A key feature of UltraTool is its independent evaluation of planning with natural language, which happens before tool usage.
  arXiv  Detail & Related papers  (2024-01-30T16:52:56Z)
• ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs [104.37772295581088]
  Open-source large language models (LLMs), e.g., LLaMA, remain significantly limited in tool-use capabilities. We introduce ToolLLM, a general tool-usetuning encompassing data construction, model training, and evaluation. We first present ToolBench, an instruction-tuning framework for tool use, which is constructed automatically using ChatGPT.
  arXiv  Detail & Related papers  (2023-07-31T15:56:53Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.