gemini-3-flash-preview

# Gemini 3 Flash Preview

This is a reasoning dataset created using **[Gemini 3 Flash Preview](https://openrouter.ai/google/gemini-3-flash-preview)** with a reasoning depth set to high.

The dataset is meant for creating distilled versions of [Gemini 3 Flash Preview](https://openrouter.ai/google/gemini-3-flash-preview) by fine-tuning already existing open-source LLMs.

This dataset contains a collection of prompts categorized by themes such as benchmarks, psychology, web development, embedded systems, creative writing, design, finance, legal, marketing, and science. No real benchmark questions were used

**No real benchmark questions were used in this dataset**, just questions generated in similar style/format and difficulty level.

*All prompts were generated by Gemini 3 Pro and GPT 5.2* - An organized list of the prompts (separated by category) can be found in the repository files.

## Usage

Load a specific config like this:

```python
from datasets import load_dataset
dataset = load_dataset("TeichAI/gemini-3-flash-preview", "pseudobench")
```

## Stats

- Costs: $ 26.44 (USD)
- Total tokens (input + output): 8.98 M

*Created via [DataGen](https://github.com/TeichAI/datagen) by TeichAI*

# Gemini 3 Flash Preview

本数据集为推理类数据集，依托**[Gemini 3 Flash Preview](https://openrouter.ai/google/gemini-3-flash-preview)** 构建，推理深度设置为高阶模式。

本数据集旨在通过对现有开源大语言模型（LLM）进行微调，以构建Gemini 3 Flash Preview的轻量化蒸馏版本。

本数据集包含多主题提示词集合，分类涵盖基准测试、心理学、Web开发、嵌入式系统、创意写作、设计、金融、法律、营销与科学领域。本数据集未使用任何真实基准测试题目，所有题目均为仿照同类风格、格式与难度级别生成的模拟题目。

*所有提示词均由Gemini 3 Pro与GPT 5.2生成* —— 按类别整理的提示词完整列表可在仓库文件中查阅。

## 使用方法
可通过如下方式加载指定配置的数据集：
python
from datasets import load_dataset
dataset = load_dataset("TeichAI/gemini-3-flash-preview", "pseudobench")

## 数据集统计
- 构建成本：26.44美元
- 总Token数（输入+输出）：8.98 M

*本数据集由TeichAI基于[DataGen](https://github.com/TeichAI/datagen)工具生成*

搜集汇总

背景与挑战

背景概述

该数据集专为通过微调现有开源大语言模型来创建Gemini 3 Flash Preview的蒸馏版本而设计。它包含按主题分类的提示词集合，这些提示词由AI生成，模拟了基准问题的风格和难度，但未使用真实基准问题。用户可通过指定配置代码加载数据集，适用于模型优化和评估场景。