Magpie-Llama-3.1-Pro-300K-Filtered


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 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-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.

**License**: Please follow [Meta Llama 3.1 Community License](https://github.com/meta-llama/llama-models/blob/main/models/llama3_1/LICENSE).

### 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)
- **Reward**: The output of the reward model given the specific instruction-response pair.
- **Language**: The language of the instruction.

## Filter Setups
- **Input Quality**: >= good
- **Instruction Reward**: >=-10
- Remove repetition and incomplete instructions (e.g., end with :)
- Choose 300K data with the longest responses

## Limitations
This dataset contains a large amount of chain-of-thought responses, which may potentially decrease the performance. Therefore, we reduce the amount of data containing `## Step 1` in the multi-turn version: [Magpie-Align/Magpie-Llama-3.1-Pro-MT-300K-v0.1](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-MT-300K-v0.1).

## Dataset Navigation 🧭
|Model Name | Dataset | Type | Description |
|-------------|:-------|:-------|:-------|
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-1M](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-1M-v0.1) | SFT | 1M Raw conversations built with Meta Llama 3.1 70B.
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-300K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-300K-Filtered) | SFT | Apply a filter and select 300K high quality conversations.
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-500K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-500K-Filtered) | SFT | Apply a filter and select 500K high quality conversations.
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-MT-500K](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-MT-500K-v0.1) | SFT | Extend Magpie-Llama-3.1-Pro-500K-Filtered to multi-turn.
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-MT-300K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-MT-300K-Filtered) | SFT | Select 300K high quality multi-turn conversations from Magpie-Llama-3.1-Pro-MT-500K.
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-DPO-100K](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-DPO-100K-v0.1) | DPO | DPO dataset via Best-of-N sampling and rewards.


项目官网：[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>点击展开</summary>
高质量指令数据对于对齐大语言模型（Large Language Model, LLM）至关重要。尽管部分模型（如 Llama-3-Instruct）开放了模型权重，但其对齐数据仍处于私有状态，这阻碍了人工智能的民主化进程。现有开源数据构建方法面临人工成本高昂、提示范围受限且预定义等问题，难以有效扩展，进而可能限制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信息，大规模合成高质量指令数据？我们提出了一种用于生成大规模对齐数据的自合成方法，命名为 Magpie。我们的核心观察是：得益于自回归特性，仅向 Llama-3-Instruct 这类已对齐大语言模型输入用户消息预留位置之前的左侧模板，模型即可生成用户查询。我们利用该方法对 Llama-3-Instruct 进行提示，生成了包含对应回复的400万条指令。我们对提取的数据进行了全面分析，并筛选出30万条高质量样本。为对比 Magpie 数据集与其他公开指令数据集的效果，我们使用每个数据集分别微调 Llama-3-8B-Base 模型，并评估微调后模型的性能。结果表明，在部分任务中，使用 Magpie 数据集微调的模型性能可与官方 Llama-3-8B-Instruct 相媲美——尽管后者通过监督微调（Supervised Fine-Tuning, SFT）及后续反馈学习使用了1000万条数据进行增强。我们还证实，仅使用 Magpie 数据集进行监督微调，其效果便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（Direct Preference Optimization, DPO）数据集。这一优势在 AlpacaEval、ArenaHard 及 WildBench 等对齐基准测试中均有体现。
</details>

## 数据集详情
本数据集由 [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) 通过 [Magpie](https://huggingface.co/Magpie-Align) 生成。有关实现细节，请参阅我们的[论文](https://arxiv.org/abs/2406.08464)与[代码仓库](https://github.com/magpie-align/magpie)。

**许可协议**：请遵循 [Meta Llama 3.1 社区许可协议](https://github.com/meta-llama/llama-models/blob/main/models/llama3_1/LICENSE)。

### 可用标签
- **输入长度**：指令的总字符数
- **输出长度**：回复的总字符数
- **任务类别**：指令所属的具体任务分类
- **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」「优秀」
- **输入难度**：完成指令描述的任务所需的知识水平，评级分为「极简单」「简单」「中等」「困难」「极困难」
- **最小邻域距离**：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
- **奖励值**：针对特定指令-回复对的奖励模型输出结果
- **语言**：指令所使用的语言

## 过滤设置
- **输入质量**：≥ 良好
- **指令奖励值**：≥ -10
- 移除重复及不完整指令（例如以冒号结尾的指令）
- 选取响应最长的30万条数据

## 局限性说明
本数据集包含大量思维链（Chain-of-Thought）回复，这可能会对模型性能产生负面影响。因此，我们在多轮版本 [Magpie-Align/Magpie-Llama-3.1-Pro-MT-300K-v0.1](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-MT-300K-v0.1) 中减少了包含`## Step 1`的样本数量。

## 数据集导航 🧭
| 模型名称 | 数据集 | 类型 | 描述 |
|-------------|:-------|:-------|:-------|
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-1M](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-1M-v0.1) | 监督微调(Supervised Fine-Tuning, SFT) | 基于 Meta Llama 3.1 70B 构建的100万条原始对话数据。
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-300K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-300K-Filtered) | 监督微调(Supervised Fine-Tuning, SFT) | 经过过滤筛选后得到的30万条高质量对话数据。
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-500K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-500K-Filtered) | 监督微调(Supervised Fine-Tuning, SFT) | 经过过滤筛选后得到的50万条高质量对话数据。
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-MT-500K](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-MT-500K-v0.1) | 监督微调(Supervised Fine-Tuning, SFT) | 将 Magpie-Llama-3.1-Pro-500K-Filtered 扩展为多轮对话格式。
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-MT-300K-Filtered](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-MT-300K-Filtered) | 监督微调(Supervised Fine-Tuning, SFT) | 从 Magpie-Llama-3.1-Pro-MT-500K 中筛选出的30万条高质量多轮对话数据。
| [Llama 3.1 70B Instruct](https://huggingface.co/meta-llama/Meta-Llama-3.1-70B-Instruct) | [Magpie-Llama-3.1-Pro-DPO-100K](https://huggingface.co/datasets/Magpie-Align/Magpie-Llama-3.1-Pro-DPO-100K-v0.1) | 直接偏好优化(Direct Preference Optimization, DPO) | 通过最佳N次采样与奖励机制构建的DPO数据集。
<be>