iJoshNh/Emo-SocraTeach-Multi

---
license: mit
language:
- vi
- en
tags:
- education
- socratic-teaching
- math
- dialogue
size_categories:
- 100K<n<1M
---

# Emo-SocraTeach-Multi

Vietnamese translation of the SocraTeach Multi-Turn dataset for Socratic teaching fine-tuning.

## Dataset Description

This dataset contains Vietnamese translations of the SocraTeach multi-turn dialogues, designed for training language models in Socratic teaching methodology.

### Data Structure

The dataset is organized as a JSON file with the following structure:

```json
{
"GSM8K_train_0": {
"question": "Vietnamese translated question...",
"analysis": "Vietnamese translated analysis...",
"answer": "72",
"steps": ["Step 1 (English)", "Step 2 (English)"],
"dialogues": {
"dialogue_id": [
{
"system": "Vietnamese system response",
"user": "Vietnamese user response",
"user_type": "(1)"
},
...
]
}
}
}
```

### Fields

- `question`: Math problem in Vietnamese
- `analysis`: Step-by-step solution analysis in Vietnamese
- `answer`: Numerical answer (not translated)
- `steps`: Problem-solving steps (kept in English)
- `dialogues`: Multi-turn Socratic teaching dialogues in Vietnamese

## Usage

```python
from datasets import load_dataset

dataset = load_dataset("iJoshNh/Emo-SocraTeach-Multi")
```

## License

MIT License