TauIndianBankBench：面向工具使用对话智能体的流程感知动作评估

AI 摘要

研究者提出 TauIndianBankBench，一个面向印度零售银行业务、评估工具使用型对话智能体的动作导向基准，包含 1000 个任务，同时检查最终记录状态和是否执行了所有必需步骤（含只读调用）。实验评估 Gemma 4 系列五个模型和 Qwen3.8-27B，发现仅检查最终记录时，每个模型有 6.5%–12.0% 的会话被误判为成功，尽管助手跳过了必需步骤或做了未被请求的更改。该工作表明需同时评估流程与结果，才能准确衡量模型在印度银行业场景的表现。

正文节选

TauIndianBankBench: Procedure-Aware, Action-Oriented Evaluation of Tool-Using Conversational Agents Abstract Many Indian banks serve customers through FAQ pages and scripted chatbots. Assistants based on language models could respond faster and carry out requests such as blocking a card, but only if they give correct information and follow the bank’s rules. These rules prescribe a procedure, a set of required steps, such as obtaining a closure quote before closing a deposit. Adapting models to t

发布时间：2026-10-24 12:00

抓取时间：2026-10-23 12:13

来源机构：arXiv