CIFQA: A Deterministic Tool-Grounded Multi-Agent LLM Framework for Financial Query Answering

CIFQA：一种用于金融问答的确定性工具驱动型多智能体大模型框架

Abstract: Calculation-intensive financial question answering requires exact reasoning over structured rates, temporal conditions, numerical formulas, and rule-based constraints. Although Large Language Models (LLMs) perform strongly on natural language tasks, they often produce numerically incorrect yet plausible answers when solving multi-step financial calculations.

摘要： 计算密集型金融问答需要对结构化利率、时间条件、数值公式和基于规则的约束进行精确推理。尽管大语言模型（LLM）在自然语言任务中表现出色，但在解决多步金融计算时，它们往往会产生数值错误但看起来合理的答案。

To address this limitation, we introduce CIFQA (Calculation-Intensive Financial Query Answering), a deterministic tool-grounded multi-agent LLM framework for financial question answering. CIFQA separates language understanding from numerical execution by assigning specialized agents to query interpretation, routing, parameter extraction, computation planning, and response generation, while deterministic Python-based tools perform financial calculations and rule application.

为了解决这一局限性，我们引入了 CIFQA（计算密集型金融问答），这是一个用于金融问答的确定性工具驱动型多智能体大模型框架。CIFQA 通过将查询解释、路由、参数提取、计算规划和响应生成分配给专门的智能体，实现了语言理解与数值执行的分离，同时利用基于 Python 的确定性工具来执行金融计算和规则应用。

We instantiate CIFQA for fixed deposit query answering and evaluate it on a curated benchmark of fixed deposit queries. CIFQA achieves 95.54% accuracy on calculation-intensive queries and 90.87% overall accuracy, substantially outperforming direct LLM baselines even when provided with complete formulas, rate cards, and benchmark instructions.

我们将 CIFQA 应用于定期存款问答，并在一个精心策划的定期存款查询基准测试集上对其进行了评估。CIFQA 在计算密集型查询中达到了 95.54% 的准确率，整体准确率达到 90.87%，即使在向直接使用的大模型基准提供完整公式、利率表和基准说明的情况下，其表现也显著优于这些基准。

Ablation studies show that deterministic components such as exact rate lookup, tenure computation, rolling-year adjustment, and premature-withdrawal logic are critical contributors to performance. Notably, a 17B open-source backbone operating within CIFQA outperforms substantially larger frontier models evaluated with the same financial information, demonstrating that architectural design is a more important determinant of numerical reliability than model scale.

消融研究表明，精确利率查询、期限计算、滚动年度调整和提前支取逻辑等确定性组件是性能提升的关键因素。值得注意的是，在 CIFQA 框架下运行的 17B 参数开源主干模型，其表现优于使用相同金融信息评估的规模大得多的前沿模型，这证明了架构设计在数值可靠性方面比模型规模更为重要。

While evaluated on fixed deposit queries, CIFQA provides a generalizable framework for calculation-intensive financial reasoning tasks.

虽然目前是在定期存款查询上进行的评估，但 CIFQA 为计算密集型金融推理任务提供了一个可推广的通用框架。