Update README.md

README.md

@@ -19,10 +19,6 @@ configs:
 
 The **30,438-example** mathematical tool-integrated reasoning SFT file used by
 the ToolElite project's [SFT-Qwen2.5-3B](https://huggingface.co/BitZone910/SFT-Qwen2.5-3B).
-It was copied from the local Tool-Star training dataset
-`toolstar_sft_math_tir.json` and renamed `sft_math_tir.json` without changing its
-contents. This is the supplied mathematical training file, not the entire
-upstream 54K mixed-domain dataset.
 
 Each record follows the Alpaca/LlamaFactory format:
 
@@ -44,11 +40,8 @@ Tool-Star file is retained; publication does not change that project's config.
 
 ## Source and license
 
-Adapted from the [Tool-Star project](https://github.com/RUC-NLPIR/Tool-Star) and its
+Adapted from the 
 [Tool-Star-SFT-54K dataset](https://huggingface.co/datasets/dongguanting/Tool-Star-SFT-54K),
 whose dataset card declares MIT. The original Tool-Star authors retain their
-attribution. ToolElite publishes this local mathematical training artifact with
-its source filename, row count, and SHA256 in `publication-manifest.json`.
+attribution. 
 
-Tool-Star paper: [Tool-Star: Empowering LLM-Brained Multi-Tool Reasoner via
-Reinforcement Learning](https://arxiv.org/abs/2505.16410).