Scalable Data Ablation Approximations for Language Models through Modular Training and Merging
Paper
• 2410.15661 • Published
• 1
Datasets and models for EMNLP paper "Scalable Data Ablation Approximations for Language Models through Modular Training and Merging"