Upload folder using huggingface_hub

README.md

@@ -8,7 +8,7 @@ English | [中文 README](README.zh_CN.md)
 
 ## 📖 Overview
 
-**Web-Bench** is a benchmark designed to evaluate the performance of LLMs in actual Web development. Web-Bench contains 50 projects, each consisting of 20 tasks with sequential dependencies. The tasks implement project features in sequence, simulating real-world human development workflows. When designing Web-Bench, we aim to cover the foundational elements of Web development: Web Standards and Web Frameworks. Given the scale and complexity of these projects, which were designed by engineers with 5-10 years of experience, each presents a significant challenge. On average, a single project takes 4–8 hours for a senior engineer to complete. 
+**Web-Bench** is a benchmark designed to evaluate the performance of LLMs in actual Web development. Web-Bench contains 50 projects, each consisting of 20 tasks with sequential dependencies. The tasks implement project features in sequence, simulating real-world human development workflows. When designing Web-Bench, we aim to cover the foundational elements of Web development: Web Standards and Web Frameworks. Given the scale and complexity of these projects, which were designed by engineers with 5-10 years of experience, each presents a significant challenge. On average, a single project takes 4–8 hours for a senior engineer to complete. On our given benchmark agent (Web-Agent), SOTA (Claude 3.7 Sonnet) achieves only 25.1\% Pass@1.
 
 The distribution of the experimental data aligns well with the current code generation capabilities of mainstream LLMs.
 ![pass@1](./docs/assets/pass-1.png)

README.zh_CN.md

@@ -8,7 +8,8 @@ license: cc-by-4.0
 
 ## **📖** Overview
 
-**Web-Bench** 是一个用于评估 LLM 在真实 Web 项目上表现的基准。Web-Bench 包含 50 个项目，每个项目包含 20 个有时序依赖关系的任务，逼真模拟了人类开发项目的过程。Web-Bench 在设计时考虑了如何覆盖 Web 应用开发所依赖的基础：Web Standards 和 Web Frameworks。由于它们的庞大规模和复杂度，以及设计项目的工程师具备 5-10 年开发经验，最终设计出来的项目对于人类资深工程师而言都具有一定的复杂度（单项目平均 4-8 小时完成）。
+**Web-Bench** 是一个用于评估 LLM 在真实 Web 项目上表现的基准。Web-Bench 包含 50 个项目，每个项目包含 20 个有时序依赖关系的任务，逼真模拟了人类开发项目的过程。Web-Bench 在设计时考虑了如何覆盖 Web 应用开发所依赖的基础：Web Standards 和 Web Frameworks。由于它们的庞大规模和复杂度，以及设计项目的工程师具备 5-10 年开发经验，最终设计出来的项目对于人类资深工程师而言都具有一定的复杂度（单项目平均 4-8 小时完成）。并且在我们给定的基准 Agent 上，SOTA（Claude 3.7 Sonnet）Pass@1 仅有 25.1%。
+
 实验数据的分布和当前主流 LLM 代码生成能力也较匹配。
 ![pass@1](./docs/assets/pass-1.png)