TICKing All the Boxes：生成式检查清单改进大语言模型评估与生成
TICKing All the Boxes: Generated Checklists Improve LLM Evaluation and Generation
AI总结：
本文提出 TICK 协议，利用 LLM 生成的指令专属检查清单实现可解释评估，显著提升 LLM 判断与人类偏好的一致性，并通过 STICK 自我精炼和 Best-of-N 选择在多个基准上改进生成质量。
摘要：
鉴于大语言模型（LLMs）的广泛采用和使用，对其指令遵循能力进行灵活且可解释的评估至关重要。模型输出之间的偏好判断已成为事实上的评估标准，尽管这种做法将复杂的、多方面的偏好压缩为单一排序。此外，由于人工标注缓慢且成本高昂，LLMs 越来越多地被用于做出这些判断，但代价是可靠性和可解释性下降。在这项工作中，我们提出了 TICK（Targeted Instruct-evaluation with ChecKlists，基于检查清单的定向指令评估），这是一种全自动、可解释的评估协议，利用 LLM 生成的、针对具体指令的检查清单来结构化评估过程。我们首先表明，给定一条指令，LLMs 能够可靠地生成高质量、量身定制的评估检查清单，将指令分解为一系列 YES/NO 问题。每个问题询问候选响应是否满足指令的某一特定要求。我们证明，与让 LLM 直接对输出打分相比，使用 TICK 使 LLM 判断与人类偏好之间完全一致的比例显著提高（46.4% → 52.2%）。随后我们展示 STICK（Self-TICK）可通过自我精炼和 Best-of-N 选择在多个基准上提升生成质量。在 LiveBench 推理任务上，STICK 自我精炼带来 +7.8% 的绝对提升，而使用 STICK 的 Best-of-N 选择在真实世界指令数据集 WildBench 上取得 +6.3% 的绝对改进。鉴于此，结构化、多方面的自我改进被证明是进一步提升 LLM 能力的有前景的途径。最后，通过向负责直接对 WildBench 指令的 LLM 响应进行打分的人类评估者提供 LLM 生成的检查清单，我们显著提高了标注者间一致性（0.194 → 0.256）。
英文摘要：
Given the widespread adoption and usage of Large Language Models (LLMs), it is crucial to have flexible and interpretable evaluations of their instruction-following ability. Preference judgments between model outputs have become the de facto evaluation standard, despite distilling complex, multi-faceted preferences into a single ranking. Furthermore, as human annotation is slow and costly, LLMs are increasingly used to make these judgments, at the expense of reliability and interpretability. In this work, we propose TICK (Targeted Instruct-evaluation with ChecKlists), a fully automated, interpretable evaluation protocol that structures evaluations with LLM-generated, instruction-specific checklists. We first show that, given an instruction, LLMs can reliably produce high-quality, tailored evaluation checklists that decompose the instruction into a series of YES/NO questions. Each question asks whether a candidate response meets a specific requirement of the instruction. We demonstrate that using TICK leads to a significant increase (46.4% $\to$ 52.2%) in the frequency of exact agreements between LLM judgements and human preferences, as compared to having an LLM directly score an output. We then show that STICK (Self-TICK) can be used to improve generation quality across multiple benchmarks via self-refinement and Best-of-N selection. STICK self-refinement on LiveBench reasoning tasks leads to an absolute gain of $+$7.8%, whilst Best-of-N selection with STICK attains $+$6.3% absolute improvement on the real-world instruction dataset, WildBench. In light of this, structured, multi-faceted self-improvement is shown to be a promising way to further advance LLM capabilities. Finally, by providing LLM-generated checklists to human evaluators tasked with directly scoring LLM responses to WildBench instructions, we notably increase inter-annotator agreement (0.194 $\to$ 0.256).
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。
