用闭环反馈生成高质量工具调用数据，提升模型工具使用能力
ToolLoop: Closed-Loop Tool-Use Data Synthesis via Decomposed Generation and Dynamic Self-Feedback
- 分三阶段生成+动态反馈，实现边生成边验证优化
- 用11K合成数据训练模型，非推理模式准确率达86.40%
- 仅用18.3%数据量，在跨基准测试中仍达72.1%准确率
高质量的工具调用数据对训练语言模型有效使用外部工具至关重要。然而，现有合成方法通常采用先生成后过滤的静态验证范式，常导致数据效率低且特征分布不均。我们提出 ToolLoop，一个闭环框架，将合成过程分解为三个渐进阶段：(1) 采样函数名组合作为真实标签；(2) 反向推导用户查询；(3) 正向推导工具调用。在每个阶段，动态自反馈迭代引导模型生成更高质量内容，实现从‘生成-过滤’到‘生成-验证-优化’的转变。在 Berkeley Function Calling Leaderboard（BFCL）上，使用11K合成样本训练的4B参数模型在非推理模式下达到86.40%准确率，而移除与BFCL重叠候选函数的Isolate变体仍达到86.07%。在ACEBench上的跨基准评估进一步证明其强泛化能力，仅使用基线18.3%的训练数据即取得72.1%的整体准确率。
原文摘要 · Abstract (English)
High-quality tool-use data is critical for training language models to interact effectively with external tools. However, existing synthetic approaches typically follow a generate-then-filter paradigm with static post-hoc verification, often yielding inefficient data with imbalanced feature distributions. We propose ToolLoop, a closed-loop framework that decomposes synthesis into three progressive stages: (1) sampling function name combinations as ground truth; (2) backward derivation of user queries; and (3) forward derivation of tool calls. At each stage, dynamic self-feedback iteratively guides the model toward high-quality generation, realizing a transition from generate-then-filter to generate-verify-refine. On the Berkeley Function Calling Leaderboard (BFCL), a 4B parameter model trained with our 11K synthetic examples achieves 86.40% accuracy in non-reasoning mode, while an Isolate variant that removes BFCL-overlapping candidate functions still reaches 86.07\%. Cross-benchmark evaluation on ACEBench further demonstrates strong generalization, with 72.1% overall accuracy using only 18.3% of baseline training data.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。