Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,39 @@
|
|
1 |
-
---
|
2 |
-
license: mit
|
3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: mit
|
3 |
+
---
|
4 |
+
# TutorQA Benchmark
|
5 |
+
|
6 |
+
This dataset is part of the benchmark introduced in the paper [Graphusion: Leveraging Large Language Models for
|
7 |
+
Scientific Knowledge Graph Fusion and Construction in NLP Education](https://arxiv.org/pdf/2407.10794v1). We also release more data in our [GitHub page](https://github.com/IreneZihuiLi/Graphusion/tree/main).
|
8 |
+
It contains 6 tasks designed for evaluating various aspects of reasoning, graph understanding, and language generation.
|
9 |
+
|
10 |
+
## Dataset Structure
|
11 |
+
|
12 |
+
Each task is a separate split:
|
13 |
+
- `task1`: Relation Judgment
|
14 |
+
- `task2`: Prerequisite Prediction
|
15 |
+
- `task3`: Path Searching
|
16 |
+
- `task4`: Subgraph Completion
|
17 |
+
- `task5`: Clustering
|
18 |
+
- `task6`: Idea Hamster (no answers, open ended)
|
19 |
+
|
20 |
+
| Split | Fields |
|
21 |
+
|:-------|:----------------------------|
|
22 |
+
| task1 | `question`, `answer` |
|
23 |
+
| task2 | `question`, `answer` |
|
24 |
+
| task3 | `question`, `answer` |
|
25 |
+
| task4 | `question`, `answer` |
|
26 |
+
| task5 | `question`, `answer` |
|
27 |
+
| task6 | `question` |
|
28 |
+
|
29 |
+
## Usage Example
|
30 |
+
|
31 |
+
```python
|
32 |
+
from datasets import load_dataset
|
33 |
+
|
34 |
+
dataset = load_dataset("li-lab/tutorqa")
|
35 |
+
|
36 |
+
# Access individual tasks
|
37 |
+
task1 = dataset["task1"]
|
38 |
+
task6 = dataset["task6"]
|
39 |
+
```
|