Update README.md

README.md

@@ -19,7 +19,9 @@ license: apache-2.0
 
 
 
-We are excited to introduce JT-Math-8B-Base: an 8-billion-parameter foundation model engineered for mathematical reasoning and the cornerstone of the JT-Math family. JT-Math-8B-Base was pre-trained on top of JT-Coder-8B-Base using an additional 210 billion tokens of high-quality mathematical and general-domain data. With a native 32,768-token context window, it provides a robust, scalable, and reproducible foundation for downstream fine-tuning—enabling researchers and developers to advance the frontier of math-centric AI applications. Technical details, training recipes, and reproducibility notes are available in our technical report.
+We are excited to introduce JT-Math-8B-Base: an 8-billion-parameter foundation model engineered for mathematical reasoning and the cornerstone of the JT-Math family. 
+JT-Math-8B-Base was pre-trained on top of JT-Coder-8B-Base using an additional 210 billion tokens of high-quality mathematical and general-domain data. 
+With a native 32,768-token context window, it provides a robust, scalable, and reproducible foundation for downstream fine-tuning—enabling researchers and developers to advance the frontier of math-centric AI applications. Technical details, training recipes, and reproducibility notes are available in our technical report.