Update README.md

README.md

@@ -22,7 +22,7 @@ size_categories:
 > It is intended **only for visualizing the data structure and format**, and for small-scale experiments.  
 > A larger / full version of the dataset may be released in future once further cleaning, de-identification, and checks are completed.  
 > 当前版本仅包含**一小部分样例数据**，用于展示数据结构与字段格式，方便他人编写和调试预处理代码。  
-> 完整或更大规模的数据集将在后续经过进一步清洗与审查后视情况逐步开放。
+> 完整或更大规模的数据集将在后续经过进一步清洗、标注核验与隐私审查后，视情况逐步开放。
 
 This repository hosts the **processed dataset** used in the paper:
 
@@ -33,7 +33,11 @@ The dataset consists of **Chinese criminal judgments** involving **multiple defe
 The dataset is designed for **judgment prediction**, **explainable legal AI**, and **multi-defendant reasoning** in Chinese criminal law.
 
 > 🇨🇳 简要中文说明：  
-> 本数据集选取了中国刑事案件中的多被告人判决书，抽取并脱敏了“事实摘要（FD）”、“裁判说理片段（CV）”以及按被告人划分的量刑、角色等标签。  
+> 本数据集来源于 **中国裁判文书网** 上公开的刑事判决书，  
+> 经过筛选后，仅保留以 **故意伤害罪（《刑法》第234条）** 为主要罪名的案件，并进一步进行了数据标注与隐私处理：  
+> - 对当事人姓名等个人身份信息进行了脱敏（如 `[张某]`、`[王某1]` 等）；  
+> - 删除或模糊处理了身份证号、联系方式、精确住址等敏感信息；  
+> - 仅保留建模所必需的字段，如事实摘要（`FD`）、裁判说理片段（`CV`）以及按被告人划分的量刑与角色标签。  
 > **当前公开版本仅为小规模样例集**，主要用于展示数据结构和字段含义，便于复现与扩展。后续可能根据数据清洗和隐私审查情况，逐步开放更大规模数据。
 
 ---
@@ -42,12 +46,12 @@ The dataset is designed for **judgment prediction**, **explainable legal AI**, a
 
 - **Code implementation** for the paper (model, training, experiments) is provided in the GitHub repository:
 
-  👉 [GitHub: *Explainable Multidefendant Judgment Prediction Enhanced by Judicial Logic Based on Large Language Models*](https://github.com/XuZhang29/MMSI)  <!-- 请改成你自己的 GitHub 链接 -->
+  👉 [GitHub: *Explainable Multidefendant Judgment Prediction Enhanced by Judicial Logic Based on Large Language Models*](https://github.com/XuZhang29/MMSI)
 
 The Hugging Face dataset `SHerZH/IMLJP` is mainly intended to:
 
-1. Provide a **public example** of the data format (`FD`, `CV`, `judgment`).
-2. Help users understand how to **construct their own datasets** in the same structure.
+1. Provide a **public example** of the data format (e.g., `FD`, `CV`, `defendants` list with per-defendant labels).
+2. Help users understand how to **construct their own datasets** in the same structure based on publicly available judgments (e.g., from China Judgments Online).
 3. Support **lightweight tests and visualization** of models and preprocessing pipelines.
 
-In future, we plan to release a **larger / full dataset** version (with the same structure) once all ethical, privacy, and legal considerations are carefully addressed.
+In future, we plan to release a **larger / full dataset** version (with the same structure and the same crime category: *intentional injury*) once all ethical, privacy, and legal considerations are carefully addressed.