README.md

---
license: apache-2.0
task_categories:
- text-generation
- question-answering
language:
- en
tags:
- reasoning
- chain-of-thought
- cot
- math
- logic
- chatml
- pinkpixel
size_categories:
- 10k<n<100k
---

# 🧠 Reasoning Mix 100k ✨

A high-quality, balanced reasoning dataset consisting of **99,999 samples** extracted from three reasoning datasets. This dataset is specifically formatted for models to utilize a `/think` block for step-by-step reasoning.

## 📊 Dataset Summary

The **Reasoning Mix 100k** is a curated collection of reasoning tasks, primarily focused on mathematics, logic, and general problem-solving. It combines the strengths of three high-performing datasets into a unified format optimized for **Chain-of-Thought (CoT)** training.

| Source | Samples | Format |
| :--- | :--- | :--- |
| **GLM-5.1-Reasoning-1M** | 33,333 | General/Math Reasoning |
| **OpenMathReasoning** | 33,333 | Mathematical Chain-of-Thought |
| **OpenThoughts-114k** | 33,333 | Comprehensive Reasoning |
| **Total** | **99,999** | **ChatML (Reasoning Optimized)** |

## 🛠️ Formatting
The dataset uses the **ChatML** format with a specialized system prompt and assistant response structure.

### 📝 System Prompt
> "You are a highly capable AI assistant. For every response, you must first show your step-by-step reasoning process within a /think block, followed by the final answer."

### 🤖 Assistant Responses
Every response is wrapped in a `/think` block structure:
```text
/think
[Step-by-step reasoning process...]
/think

[Final answer or solution]
```

## 📂 Dataset Structure

Each sample in the `reasoning_mix_100k.jsonl` file follows this structure:

```json
{
  "messages": [
    {
      "role": "system",
      "content": "You are a highly capable AI assistant..."
    },
    {
      "role": "user",
      "content": "Evaluate the integral..."
    },
    {
      "role": "assistant",
      "content": "/think\n[Reasoning...]\n/think\n\n[Final Answer]"
    }
  ]
}
```

## 🚀 Usage

You can load this dataset using the Hugging Face `datasets` library:

```python
from datasets import load_dataset

dataset = load_dataset("pinkpixel/reasoning-mix-100k")
print(dataset["train"][0])
```

## 📜 License

This dataset is released under the **Apache License 2.0**.

---
Made with ❤️ by **Pink Pixel** ✨