pw-ai-research commited on
Commit
6479eec
·
verified ·
1 Parent(s): ac9b78a

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +87 -3
README.md CHANGED
@@ -1,3 +1,87 @@
1
- ---
2
- license: apache-2.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ ---
4
+ # JEE Mains 2025 Math Evaluation Set
5
+
6
+ ## 🧾 Dataset Summary
7
+
8
+ This dataset contains **475 math questions** from the **official JEE Mains 2025** examination, covering both **January** and **April** sessions. It is curated to benchmark mathematical reasoning models under high-stakes exam conditions.
9
+
10
+ ---
11
+
12
+ ## 🚀 How to Load the Dataset
13
+
14
+ You can load the evaluation data using the `datasets` library from Hugging Face:
15
+
16
+ ```python
17
+ from datasets import load_dataset
18
+
19
+ # Load January session evaluation set
20
+ jan_data = load_dataset("PhysicsWallahAI/JEE-Main-2025-Math", "jan", split="test")
21
+
22
+ # Load April session evaluation set
23
+ apr_data = load_dataset("PhysicsWallahAI/JEE-Main-2025-Math", "apr", split="test")
24
+ ```
25
+
26
+ ---
27
+
28
+ ## 📂 Dataset Structure
29
+
30
+ Each sample is stored as a JSON object with the following fields:
31
+
32
+ | Field Name | Type | Description |
33
+ |--------------------|-----------|-----------------------------------------------------------------------------|
34
+ | `question` | `string` | Math problem text (can include LaTeX) |
35
+ | `answer` | `string` | Final answer (NAT or symbolic form) |
36
+ | `question_type` | `int` | `0 = Numerical Answer Type`, `1 = Multiple Choice Question` |
37
+ | `options` | `list` | List of answer choices (present only for MCQ) |
38
+ | `correct_options` | `list` | Indices of correct options in `options[]` (for MCQ only) |
39
+ | `additional_data` | `dict` | Placeholder for extended fields used during model training or evaluation. |
40
+ | `metadata` | `dict` | Optional metadata providing contextual information about the question. |
41
+
42
+ ---
43
+
44
+ ## 📊 Dataset Statistics
45
+
46
+ | Split | Papers | Questions |
47
+ | ------------ | ------ | --------- |
48
+ | January 2025 | 10 | 250 |
49
+ | April 2025 | 9 | 225 |
50
+ | **Total** | 19 | **475** |
51
+
52
+ * **MCQs**: \~80%
53
+ * **NATs**: \~20%
54
+
55
+ ---
56
+
57
+ ## 📥 Source
58
+
59
+ All questions were sourced from **official JEE Mains 2025** mathematics papers publicly released by **NTA**. Answer keys were cross-verified with NTA final answer releases.
60
+
61
+ ---
62
+
63
+ ## 💼 Intended Uses
64
+
65
+ * Benchmarking Indian math LLMs
66
+ * Evaluating symbolic + numeric reasoning
67
+ * Comparing SFT/RLHF/retrieval-based models on real exams
68
+
69
+ ---
70
+
71
+ ## ⚠️ Limitations
72
+
73
+ * Limited to mathematics domain
74
+
75
+ ---
76
+
77
+ ## 📄 Citation
78
+
79
+ ```bibtex
80
+ @misc{jee2025math,
81
+ title = {JEE Mains 2025 Math Evaluation Set},
82
+ author = {Physics Wallah AI Research},
83
+ year = {2025},
84
+ note = {Official JEE Mains 2025 math questions curated for evaluating educational language models},
85
+ howpublished = {\url{https://huggingface.co/datasets/PhysicsWallahAI/JEE-Main-2025-Math}},
86
+ }
87
+ ```