Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -100,23 +100,19 @@ Each example includes:
|
|
| 100 |
],
|
| 101 |
"conceptual_takeaway": "..."
|
| 102 |
}
|
|
|
|
| 103 |
|
| 104 |
-
Intended use
|
| 105 |
|
| 106 |
This dataset is suitable for:
|
| 107 |
|
| 108 |
-
supervised fine-tuning on step-by-step mathematical reasoning
|
|
|
|
|
|
|
|
|
|
|
|
|
| 109 |
|
| 110 |
-
|
| 111 |
-
|
| 112 |
-
answer verification and self-check training
|
| 113 |
-
|
| 114 |
-
error detection and correction tasks
|
| 115 |
-
|
| 116 |
-
curriculum-style difficulty experiments
|
| 117 |
-
|
| 118 |
-
|
| 119 |
-
Notes
|
| 120 |
|
| 121 |
The dataset is synthetic and intentionally structured for reasoning quality.
|
| 122 |
|
|
@@ -126,8 +122,7 @@ The error catalogue is designed to model common student and model mistakes.
|
|
| 126 |
|
| 127 |
The prompts are intentionally consistent in style to make reasoning supervision easier.
|
| 128 |
|
| 129 |
-
|
| 130 |
-
Limitations
|
| 131 |
|
| 132 |
The dataset is not a collection of real-world student work.
|
| 133 |
|
|
@@ -135,8 +130,7 @@ The writing style is intentionally template-like, which may reduce natural-langu
|
|
| 135 |
|
| 136 |
The dataset focuses on mathematical reasoning and does not aim to cover general open-domain QA.
|
| 137 |
|
| 138 |
-
|
| 139 |
-
Citation
|
| 140 |
|
| 141 |
If you use this dataset, please cite it as:
|
| 142 |
|
|
|
|
| 100 |
],
|
| 101 |
"conceptual_takeaway": "..."
|
| 102 |
}
|
| 103 |
+
```
|
| 104 |
|
| 105 |
+
## Intended use
|
| 106 |
|
| 107 |
This dataset is suitable for:
|
| 108 |
|
| 109 |
+
- supervised fine-tuning on step-by-step mathematical reasoning
|
| 110 |
+
- multi-solution reasoning behavior
|
| 111 |
+
- answer verification and self-check training
|
| 112 |
+
- error detection and correction tasks
|
| 113 |
+
- curriculum-style difficulty experiments
|
| 114 |
|
| 115 |
+
## Notes
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 116 |
|
| 117 |
The dataset is synthetic and intentionally structured for reasoning quality.
|
| 118 |
|
|
|
|
| 122 |
|
| 123 |
The prompts are intentionally consistent in style to make reasoning supervision easier.
|
| 124 |
|
| 125 |
+
## Limitations
|
|
|
|
| 126 |
|
| 127 |
The dataset is not a collection of real-world student work.
|
| 128 |
|
|
|
|
| 130 |
|
| 131 |
The dataset focuses on mathematical reasoning and does not aim to cover general open-domain QA.
|
| 132 |
|
| 133 |
+
## Citation
|
|
|
|
| 134 |
|
| 135 |
If you use this dataset, please cite it as:
|
| 136 |
|