Update README: add What's New section, unify pp notation for scores

README.md

@@ -45,6 +45,10 @@ default_config_name: UltraData-Math-L3-Conversation-Synthetic
 
 ***UltraData-Math*** is a large-scale, high-quality mathematical pre-training dataset totaling **290B+ tokens** across three progressive tiers—**L1** (170.5B tokens web corpus), **L2** (33.7B tokens quality-selected), and **L3** (88B tokens multi-format refined)—designed to systematically enhance mathematical reasoning in LLMs. It has been applied to the mathematical pre-training of the [MiniCPM Series](https://huggingface.co/collections/openbmb/minicpm4) models.
 
+## 🆕 What's New
+
+- **2026.02.09**: Released UltraData-Math (290B+ tokens), a large-scale high-quality mathematical pre-training dataset with three progressive tiers (L1/L2/L3).
+
 ## 📚 Introduction
 
 High-quality pre-training data is crucial for enhancing the mathematical reasoning capabilities of large language models (LLMs). However, existing mathematical pre-training data construction schemes have the following shortcomings:
@@ -60,7 +64,7 @@ To address these issues, we propose ***UltraData-Math***—a large-scale high-qu
 - **L2 Selected Data**: Uses proprietary large models to annotate seed data and distills it into a lightweight embedding classifier to achieve efficient quality grading of the full corpus.
 - **L3 Refined Data**: Produces structured content with clear reasoning through rewriting, synthetic generation, and refinement in various formats such as Q&A, multi-turn dialogues, multi-style rewriting, and knowledge-grounded textbooks.
 
-Experiments show that on the MiniCPM-1.2B architecture, ***UltraData-Math*** achieves a score of **37.02** on the MATH500 benchmark, an improvement of **+3.62 pp** compared to Nemotron-CC 4plus; it achieves **61.79** on GSM8K, an improvement of **+3.34 pp**, while maintaining code generation and general knowledge capabilities.
+Experiments show that on the MiniCPM-1.2B architecture, ***UltraData-Math*** achieves a score of **37.02pp** on the MATH500 benchmark, an improvement of **+3.62pp** compared to Nemotron-CC 4plus; it achieves **61.79pp** on GSM8K, an improvement of **+3.34pp**, while maintaining code generation and general knowledge capabilities.
 
 ***UltraData-Math*** has been applied to the mathematical pre-training of the [MiniCPM Series](https://huggingface.co/collections/openbmb/minicpm-4-6841ab29d180257e940baa9b) models.
 

README_ZH.md

@@ -25,7 +25,7 @@
 - **L2 精选数据层**：使用闭源大模型标注种子数据并蒸馏至轻量 embedding 分类器，实现全量语料的高效质量分级。
 - **L3 精炼数据层**：通过改写、合成生成与精炼，生成具有清晰推理链条的结构化内容，涵盖 Q&A、多轮对话、多风格改写、知识教材等多种格式。
 
-实验表明，在 MiniCPM-1.2B 架构上，***UltraData-Math*** 在 MATH500 基准上达到 **37.02** 分，相较 Nemotron-CC 4plus 提升 **+3.62 pp**；在 GSM8K 上达到 **61.79** 分，提升 **+3.34 pp**，同时保持代码生成与通用知识能力。
+实验表明，在 MiniCPM-1.2B 架构上，***UltraData-Math*** 在 MATH500 基准上达到 **37.02pp**，相较 Nemotron-CC 4plus 提升 **+3.62pp**；在 GSM8K 上达到 **61.79pp**，提升 **+3.34pp**，同时保持代码生成与通用知识能力。
 
 ***UltraData-Math*** 已应用于 [MiniCPM 系列](https://huggingface.co/collections/openbmb/minicpm-4-6841ab29d180257e940baa9b) 模型的数学预训练。