Complete Recipe
"Complete Recipe" research focuses on developing efficient and effective methods for training and improving large language models (LLMs) and other machine learning models. Current research emphasizes techniques like programmatic data generation, in-context reinforcement learning, and innovative training strategies (e.g., continued pretraining, loss function modifications) to enhance model performance, particularly in handling long contexts and diverse data. These advancements are significant because they address the high computational cost and data limitations associated with training powerful models, leading to more accessible and efficient AI solutions across various applications.
Papers
Cookbook: A framework for improving LLM generative abilities via programmatic data generating templates
Avanika Narayan, Mayee F. Chen, Kush Bhatia, Christopher Ré
A Recipe For Building a Compliant Real Estate Chatbot
Navid Madani, Anusha Bagalkotkar, Supriya Anand, Gabriel Arnson, Rohini Srihari, Kenneth Joseph