1 dataset found

Tags: continual pre-training

Filter Results

  • Proof-Pile-2

    The dataset used for continual pre-training of large language models, with a focus on balancing the text distribution and mitigating overfitting.
You can also access this registry using the API (see API Docs).