Update README: Simplify benchmark list and unify table headers

README.md

@@ -132,9 +132,9 @@ Natural web data is mostly declarative text. To enhance the model's instruction
 
 We used the **MiniCPM-1.2B** model architecture and **MiniCPM3-4B** tokenizer for experimental verification. Each experiment was conducted with a training volume of **100 billion Tokens**, using the **Decay Verification** method (annealing from a 1.3T base model). We used the Lighteval library for model evaluation. Evaluation benchmarks include:
 
-- **Mathematical Reasoning:** GSM8K (4-shot), MATH (4-shot), Math-Bench
-- **Code Generation:** HumanEval (0-shot), MBPP (3-shot)
-- **Comprehensive Knowledge:** MMLU (5-shot), MMLU-STEM (5-shot)
+- **Mathematical Reasoning:** GSM8K, MATH, Math-Bench, R-Bench-Math
+- **Code Generation:** HumanEval, MBPP
+- **Comprehensive Knowledge:** MMLU, MMLU-STEM
 
 ### 🔧 Experimental Setup
 
@@ -169,7 +169,7 @@ To validate the effectiveness of our L0-L3 hierarchical framework, we conducted
 
 We used a single dataset for independent training to directly compare the effects of different data sources:
 
-| Model | Average | MMLU | MMLU-STEM | Math | GSM8K | MBPP | HumanEval | R-bench | Math-bench |
+| Model | Average | MMLU | MMLU-STEM | Math | GSM8K | MBPP | HumanEval | R-Bench-Math | Math-Bench |
 |:---|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|
 | **UltraData-Math (Ours)** | **43.79** | 51.67 | 45.93 | **37.02** | **61.79** | **49.27** | 32.93 | 23.38 | **48.33** |
 | Nemotron-cc 4plus mind | 43.45 | 52.09 | 45.99 | 35.96 | 59.97 | 48.03 | 34.76 | **23.51** | 47.25 |