Llama-3-Magpie-Pro-1M-v0.1


Project Web: [https://magpie-align.github.io/](https://magpie-align.github.io/)

Arxiv Technical Report: [https://arxiv.org/abs/2406.08464](https://arxiv.org/abs/2406.08464)

Codes: [https://github.com/magpie-align/magpie](https://github.com/magpie-align/magpie)

## Abstract
<details><summary>Click Here</summary>
High-quality instruction data is critical for aligning large language models (LLMs). Although some models, such as Llama-3-Instruct, have open weights, their alignment data remain private, which hinders the democratization of AI. High human labor costs and a limited, predefined scope for prompting prevent existing open-source data creation methods from scaling effectively, potentially limiting the diversity and quality of public alignment datasets. Is it possible to synthesize high-quality instruction data at scale by extracting it directly from an aligned LLM? We present a self-synthesis method for generating large-scale alignment data named Magpie. Our key observation is that aligned LLMs like Llama-3-Instruct can generate a user query when we input only the left-side templates up to the position reserved for user messages, thanks to their auto-regressive nature. We use this method to prompt Llama-3-Instruct and generate 4 million instructions along with their corresponding responses. We perform a comprehensive analysis of the extracted data and select 300K high-quality instances. To compare Magpie data with other public instruction datasets, we fine-tune Llama-3-8B-Base with each dataset and evaluate the performance of the fine-tuned models. Our results indicate that in some tasks, models fine-tuned with Magpie perform comparably to the official Llama-3-8B-Instruct, despite the latter being enhanced with 10 million data points through supervised fine-tuning (SFT) and subsequent feedback learning. We also show that using Magpie solely for SFT can surpass the performance of previous public datasets utilized for both SFT and preference optimization, such as direct preference optimization with UltraFeedback. This advantage is evident on alignment benchmarks such as AlpacaEval, ArenaHard, and WildBench.
</details><be>

## Dataset Details
This dataset is generated by [Llama 3 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) using [Magpie](https://huggingface.co/Magpie-Align). Please refer to our [paper](https://arxiv.org/abs/2406.08464) and [codebase](https://github.com/magpie-align/magpie) for implementation details.

This is the raw data. Feel free to apply your own filter!

### Available Labels
- **Input Length**: The total number of characters in the instructions.
- **Output Length**: The total number of characters in the responses.
- **Task Category**: The specific category of the instructions.
- **Input Quality**: The clarity, specificity, and coherence of the instructions, rated as 'very poor', 'poor', 'average', 'good', and 'excellent'.
- **Input Difficulty**: The level of knowledge required to address the task described in the instruction, rated as 'very easy', 'easy', 'medium', 'hard', or 'very hard'.
- **Minimum Neighbor Distance**: The embedding distance to the nearest neighbor within the dataset. It can be used for filtering out repetitive or similar instances.
- **Safety**: Safety tags marked by [meta-llama/Meta-Llama-Guard-2-8B](https://huggingface.co/meta-llama/Meta-Llama-Guard-2-8B)
- **Instruct Reward**: The output of the reward model given the specific instruction-response pair.
- **Base Reward**: The output of the reward model given the instruction and response from the base model.
- **Reward Difference**: Instruct Reward - Base Reward.

## Dataset Navigation 🧭
|Model Name | Dataset | Type | Description |
|-------------|:-------|:-------|:-------|
| [Llama 3 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) | [Magpie-Pro-1M](https://huggingface.co/datasets/Magpie-Align/Llama-3-Magpie-Pro-1M-v0.1) | SFT | 1M Raw conversations built with Meta Llama 3 70B.
| [Llama 3 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) | [Magpie-Pro-300K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Pro-300K-Filtered) | SFT | Apply a filter and select 300K high quality conversations.
| [Llama 3 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) | [Magpie-Pro-MT-300K](https://huggingface.co/datasets/Magpie-Align/Magpie-Pro-MT-300K-v0.1) | SFT | Select 300K difficult questions and extend to multi-turn conversations.
| [Llama 3 8B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) | [Magpie-Air-3M](https://huggingface.co/datasets/Magpie-Align/Llama-3-Magpie-Air-3M-v0.1) | SFT | 3M Raw conversations built with Meta Llama 3 8B.
| [Llama 3 8B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) | [Magpie-Air-300K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Air-300K-Filtered) | SFT | Apply a filter and select 300K high quality data.
| [Llama 3 8B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) | [Magpie-Air-MT-300K](https://huggingface.co/datasets/Magpie-Align/Magpie-Air-MT-300K-v0.1) | SFT | Select 300K difficult questions and extend to multi-turn conversations.


项目主页：[https://magpie-align.github.io/](https://magpie-align.github.io/)

Arxiv技术报告：[https://arxiv.org/abs/2406.08464](https://arxiv.org/abs/2406.08464)

代码仓库：[https://github.com/magpie-align/magpie](https://github.com/magpie-align/magpie)

## 摘要
<details><summary>点击此处展开</summary>
高质量的指令数据对于大语言模型（Large Language Model，LLM）的对齐至关重要。尽管诸如Llama-3-Instruct等部分模型已开放权重，但其对齐数据仍处于保密状态，这阻碍了人工智能的民主化进程。高昂的人力成本与受限的预定义提示范围，使得现有开源数据构建方法难以有效规模化扩展，进而可能制约了公开对齐数据集的多样性与质量。能否直接从已对齐的大语言模型中提取数据，以此规模化合成高质量的指令数据？我们提出了一种名为Magpie的自合成方法，用于生成大规模对齐数据。我们的核心观察在于：得益于自回归特性，诸如Llama-3-Instruct这类已对齐的大语言模型，仅需输入至用户消息预留位置的左侧模板，即可生成用户查询。我们借助该方法对Llama-3-Instruct进行提示，生成了400万条指令及其对应的回复。我们对提取得到的数据开展了全面分析，并从中筛选出30万条高质量样本。为了将Magpie数据集与其他公开指令数据集进行对比，我们分别使用各数据集对Llama-3-8B-Base进行微调，并评估微调后模型的性能。我们的实验结果表明，在部分任务中，使用Magpie数据集微调得到的模型性能可与官方Llama-3-8B-Instruct相媲美——尽管后者通过1000万条数据的监督微调（Supervised Fine-Tuning，SFT）与后续反馈学习完成了性能增强。此外我们还发现，仅使用Magpie数据集开展监督微调，其性能可超越此前同时用于监督微调与偏好优化的公开数据集，例如结合UltraFeedback的直接偏好优化方法。该优势在AlpacaEval、ArenaHard与WildBench等对齐基准测试中均有体现。
</details><be>

## 数据集详情
本数据集由[Llama 3 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct)借助[Magpie](https://huggingface.co/Magpie-Align)生成。如需了解实现细节，请参考我们的[论文](https://arxiv.org/abs/2406.08464)与[代码仓库](https://github.com/magpie-align/magpie)。

此为原始数据，您可自行进行筛选处理！

### 可用标签
- **输入长度**：指令中的总字符数。
- **输出长度**：回复中的总字符数。
- **任务类别**：指令所属的具体类别。
- **输入质量**：指令的清晰度、特异性与连贯性，评级分为「极差」「差」「一般」「良好」与「优秀」。
- **输入难度**：完成指令描述任务所需的知识水平，评级分为「极简单」「简单」「中等」「困难」与「极困难」。
- **最小近邻距离**：数据集中与当前样本最近邻的嵌入距离，可用于过滤重复或相似样本。
- **安全性**：由[meta-llama/Meta-Llama-Guard-2-8B](https://huggingface.co/meta-llama/Meta-Llama-Guard-2-8B)标记的安全标签。
- **指令奖励分**：针对特定指令-回复对的奖励模型输出结果。
- **基座模型奖励分**：针对基座模型生成的指令与回复的奖励模型输出结果。
- **奖励分差值**：指令奖励分与基座模型奖励分的差值。

## 数据集导航 🧭
|模型名称 | 数据集 | 类型 | 描述 |
|:-------------|:-------|:-------|:-------|
| [Llama 3 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) | [Magpie-Pro-1M](https://huggingface.co/datasets/Magpie-Align/Llama-3-Magpie-Pro-1M-v0.1) | 监督微调（SFT） | 由Meta Llama 3 70B构建的100万条原始对话数据。
| [Llama 3 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) | [Magpie-Pro-300K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Pro-300K-Filtered) | 监督微调（SFT） | 经过筛选后选出的30万条高质量对话数据。
| [Llama 3 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-70B-Instruct) | [Magpie-Pro-MT-300K](https://huggingface.co/datasets/Magpie-Align/Magpie-Pro-MT-300K-v0.1) | 监督微调（SFT） | 选取30万条高难度问题并扩展为多轮对话数据。
| [Llama 3 8B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) | [Magpie-Air-3M](https://huggingface.co/datasets/Magpie-Align/Llama-3-Magpie-Air-3M-v0.1) | 监督微调（SFT） | 由Meta Llama 3 8B构建的300万条原始对话数据。
| [Llama 3 8B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) | [Magpie-Air-300K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Air-300K-Filtered) | 监督微调（SFT） | 经过筛选后选出的30万条高质量数据。
| [Llama 3 8B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) | [Magpie-Air-MT-300K](https://huggingface.co/datasets/Magpie-Align/Magpie-Air-MT-300K-v0.1) | 监督微调（SFT） | 选取30万条高难度问题并扩展为多轮对话数据。