让大模型工作流自动满足类型约束，提升复杂任务准确性。
Type-Compliant Adaptation Cascades: Adapting Programmatic LM Workflows to Data
- 将工作流视为带类型的概率程序，用梯度法优化
- 在多个结构化任务上准确率最高提升近3倍
- 适合需要严格逻辑合规的自动化系统开发者
可靠地组合大语言模型（LLMs）以完成复杂多步任务仍是重大挑战。主流方法——优化离散提示的流水线——极易失效，难以保证结构化任务所需的正式合规性。我们提出类型合规适配级联（TACs），将工作流适配重构为学习带类型的概率程序。TACs将整个工作流（由参数高效适配的LLM和确定性逻辑构成）视为一个未归一化的联合分布，即使存在隐含中间结构，也能进行原则性的梯度训练。我们提供了可计算优化目标的理论依据，证明当模型学会类型合规时，优化偏差趋于消失。实证表明，TACs显著优于当前最优的提示优化基线。在结构化任务上表现尤为突出：对于Qwen 3 8B模型，FinQA准确率从12.0%提升至24.7%；Gemma 2 27B模型在MGSM-SymPy任务上从57.1%升至75.9%；MGSM从1.6%提升至27.3%；Gemma 7B模型在MuSR任务上从36.5%增至62.6%。TACs为构建可靠、任务合规的LLM系统提供了稳健且理论坚实的新范式。
原文摘要 · Abstract (English)
Reliably composing Large Language Models (LLMs) for complex, multi-step workflows remains a significant challenge. The dominant paradigm -- optimizing discrete prompts in a pipeline -- is notoriously brittle and struggles to enforce the formal compliance required for structured tasks. We introduce Type-Compliant Adaptation Cascades (TACs), a framework that recasts workflow adaptation as learning typed probabilistic programs. TACs treat the entire workflow, which is composed of parameter-efficiently adapted LLMs and deterministic logic, as an unnormalized joint distribution. This enables principled, gradient-based training even with latent intermediate structures. We provide theoretical justification for our tractable optimization objective, proving that the optimization bias vanishes as the model learns type compliance. Empirically, TACs significantly outperform state-of-the-art prompt-optimization baselines. Gains are particularly pronounced on structured tasks, improving FinQA from $12.0\%$ to $24.7\%$ for a Qwen 3 8B model, MGSM-SymPy from $57.1\%$ to $75.9\%$ for a Gemma 2 27B model, MGSM from $1.6\%$ to $27.3\%$, and MuSR from $36.5\%$ to $62.6\%$ for a Gemma 7B model. TACs offer a robust and theoretically grounded paradigm for developing reliable, task-compliant LLM systems.
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。