新闻
PPL-Factory: Task-Aware and Budget-Aware Data Selection from Language Modeling to Reasoning
arXiv cs.AI · 发布于 · 阅读约3分钟
30秒读懂
- 发生了什么
- PPL-Factory selects training data for fine-tuning language models using task-aware perplexity scores and budget constraints, achieving full-data accuracy with only 10% of samples.
- 为何重要
- Engineers fine-tuning large language models on reasoning tasks like math problems need to reduce computational costs without sacrificing downstream performance.
- 注意
- Results demonstrated on GSM8K and MATH datasets; effectiveness on other task domains and model sizes remains unclear from this paper.
- language model
- reasoning
- fine-tun
这条新闻背后的模式
每个模式都讲清楚技术如何运作、何时值得投入，以及在哪里会失效。
The Agent Architect
每周一个模式、一个权衡、一个生产事故案例。为构建智能体系统的人准备的每周简报。
每周一封邮件,一键退订。您的地址仅用于发送简报。