Gemma 2: Improving Open Language Models at a Practical Size
This work provides incremental improvements to lightweight open language models, benefiting the AI community by offering efficient and competitive alternatives for practical applications.
The paper tackles the problem of improving open language models at practical sizes by introducing Gemma 2, which applies known technical modifications and knowledge distillation to achieve state-of-the-art performance for models ranging from 2B to 27B parameters, offering competitive alternatives to models 2-3 times larger.
In this work, we introduce Gemma 2, a new addition to the Gemma family of lightweight, state-of-the-art open models, ranging in scale from 2 billion to 27 billion parameters. In this new version, we apply several known technical modifications to the Transformer architecture, such as interleaving local-global attentions (Beltagy et al., 2020a) and group-query attention (Ainslie et al., 2023). We also train the 2B and 9B models with knowledge distillation (Hinton et al., 2015) instead of next token prediction. The resulting models deliver the best performance for their size, and even offer competitive alternatives to models that are 2-3 times bigger. We release all our models to the community.