You Only Edit Once： 通过局部示例精修激发 LLM 的上下文能力 挑选最佳少样本示例是一种缓慢的 System-2 搜索：组合爆炸，且往往需要反复调用 LLM。 我们把它变成了 System 1。⚡ Jev-LDE，一个 1.7B 的编辑器，扫一眼检索到的示例，只做一次编辑。LLM 只回答一次。 平均 1-shot 准确率 81.2 → 88.1 You Only Edit Once 🧵 动画演示（示意示例）。查询：“How far is it from Denver to Aspen?” 语义 TopK 检索出三个相似示例：“Where is Aspen, Colorado?”（Location）、“What state is Denver in?”（Location）、“Who founded Denver?”（Person）。Jev-LDE，一个 1.7B 的 System-1 编辑器，标记出第一个虽然主题相同但答案类型错误，并输出一个动作：将 S1 替换为候选 C1，“How far is Boston from NYC?”（Number）。冻结的目标 LLM 随后回答“Number”，这是正确的；若不编辑，它会回答“Location”。结尾卡片：在 3 个基准和 4 个目标 LLM 上，平均 1-shot 准确率从 81.2 提升至 88.1，在 48 个设置中有 44 个达到最佳或并列最佳，墙钟时间增加 11%。
You Only Edit Once:
Incentivizing In-Context Capability of LLMs via Local Demonstration Refinement
You Only Edit Once: Incentivizing In-Context Capability of LLMs via Local Demonstration Refinement Picking the best few-shot demos is a slow System-2 search: combinatorial, often with repeated LLM calls. We made it System 1. ⚡ Jev-LDE, a 1.7B editor, glances at the retrieved demos and makes ONE edit. The LLM answers once. Avg 1-shot acc 81.2 → 88.1 You Only Edit Once 🧵 Animated walkthrough (illustrative example). Query: "How far is it from Denver to Aspen?" Semantic TopK retrieves three look-alike demos: "Where is Aspen, Colorado?" (Location), "What state is Denver in?" (Location), "Who founded Denver?" (Person). Jev-LDE, a 1.7B System-1 editor, flags the first as same topic but wrong answer type and outputs one action: Replace S1 with candidate C1, "How far is Boston from NYC?" (Number). The frozen target LLM then answers "Number", which is correct; without the edit it answers "Location". End card: average 1-shot accuracy 81.2 to 88.1 across 3 benchmarks and 4 target LLMs, best or tied-best in 44 of 48 settings, +11% wall time.在 X 查看被引用的帖子
来源：Dongxi 东锡 NLP · x.com