Chain-of-Instructions：面向大语言模型的组合式指令微调
Chain-of-Instructions: Compositional Instruction Tuning on Large Language Models
AI总结：
本文提出 chain-of-instructions（CoI）组合式指令及 CoI-tuning，通过让模型逐步完成相互衔接的子任务，提升 LLM 对复杂、未见和长指令链的泛化能力。
摘要：
使用大量多样化指令对大语言模型（LLM）进行微调，已提升模型对不同任务的泛化能力，甚至包括未见过的任务。然而，现有大多数指令数据集只包含单一指令，模型在遵循由多个子任务组成的复杂指令时存在困难。在本研究中，我们提出了一个名为指令链（chain-of-instructions，CoI）的组合式指令新概念，其中一条指令的输出像链条一样成为下一条指令的输入。与解决单一指令任务的传统做法不同，我们提出的方法鼓励模型逐步解决每个子任务，直到得到最终答案。CoI-tuning（即使用 CoI 指令进行微调）提升了模型处理由多个子任务组成的指令以及未见过的组合任务（如多语言摘要）的能力。总体而言，我们的研究发现，对现有指令数据进行简单的 CoI 微调，即可带来持续的泛化能力，使模型能够解决更复杂、未见过且更长的指令链。
英文摘要：
Fine-tuning large language models (LLMs) with a collection of large and diverse instructions has improved the model's generalization to different tasks, even for unseen tasks. However, most existing instruction datasets include only single instructions, and they struggle to follow complex instructions composed of multiple subtasks. In this work, we propose a novel concept of compositional instructions called chain-of-instructions (CoI), where the output of one instruction becomes an input for the next like a chain. Unlike the conventional practice of solving single instruction tasks, our proposed method encourages a model to solve each subtask step by step until the final answer is reached. CoI-tuning (i.e., fine-tuning with CoI instructions) improves the model's ability to handle instructions composed of multiple subtasks as well as unseen composite tasks such as multilingual summarization. Overall, our study find that simple CoI tuning of existing instruction data can provide consistent generalization to solve more complex, unseen, and longer chains of instructions.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
