Summary of Small Language Models: Survey, Measurements, and Insights, by Zhenyan Lu et al.
Small Language Models: Survey, Measurements, and Insights
by Zhenyan Lu, Xiang Li, Dongqi Cai, Rongjie Yi, Fangming Liu, Xiwen Zhang, Nicholas D. Lane, Mengwei Xu
First submitted to arxiv on: 24 Sep 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper explores the development of small language models (SLMs), which are used in modern smart devices to make machine intelligence more accessible, affordable, and efficient for everyday tasks. The authors analyze 70 state-of-the-art open-source SLMs, focusing on transformer-based, decoder-only language models with 100M-5B parameters. They evaluate the models’ capabilities in various domains, including commonsense reasoning, mathematics, in-context learning, and long context. Additionally, they benchmark their inference latency and memory footprints to gain insights into their on-device runtime costs. The authors hope that this research will advance the field of SLMs. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about small language models (SLMs) that are used in smart devices like phones and computers. These models help make machine intelligence more accessible and efficient for everyday tasks. The researchers looked at 70 different SLMS to see how they work and what they’re good at. They tested the models’ abilities in areas like math, common sense, and learning from context. They also measured how fast these models run on devices and how much memory they need. The goal is to make SLMs better and more useful for people. |
Keywords
» Artificial intelligence » Decoder » Inference » Transformer
