instinct-data

# Instinct Dataset This repository contains the next-edit data used to train and evaluate Continue's state-of-the-art open Next Edit model, [Instinct](https://huggingface.co/continuedev/instinct). The splits are given by language, with Typescript being the original, and other languages bootstrapped synthetically off of the Typescript data. For more information on the dataset, please refer to our [blog post](https://blog.continue.dev/instinct/). We additionally have code available on [GitHub](https://github.com/continuedev/instinct).# Instinct 数据集 本仓库收录了用于训练与评估Continue团队研发的当前最优开源下一轮编辑（Next Edit）模型[Instinct](https://huggingface.co/continuedev/instinct)的下一轮编辑数据集。该数据集按语言维度划分，其中TypeScript为原始基准语言，其余语言的数据集均基于TypeScript数据经合成引导构建。如需了解该数据集的更多细节，请参阅我们的[博客文章](https://blog.continue.dev/instinct/)。相关代码已开源至[GitHub](https://github.com/continuedev/instinct)

搜集汇总

背景与挑战

背景概述

该数据集由continuedev开发，用于训练和评估Instinct模型，包含基于TypeScript生成的多语言编辑数据。它采用Apache-2.0许可证，更新于2025年9月6日，体积约为50.62MB。