Published Jun 21, 2024
Frontiers in synthetic data
Join Nathan Lambert as he delves into the revolutionary role of synthetic data in language model development, examining iterative filtering techniques to prevent model collapse and enhance strategies, while discussing the potential and challenges of licensing and usage in multi-output datasets.
Topics covered
Popular Clips
Episode Highlights
Related Episodes
A recipe for frontier model post-training
Answers 383 questionsWe aren't running out of training data, we are running out of open training data
Answers 383 questionsFutures of the data foundry business model
Answers 383 questionsA realistic path to robotic foundation models
Answers 383 questionsSwitched to Claude 3.5
Answers 383 questions
Interviewing Riley Goodside on the science of prompting
Answers 383 questionsNous Hermes 3 and exploiting underspecified evaluations
Answers 383 questionsText-to-video AI is already abundant
Answers 383 questionsModel merging lessons in The Waifu Research Department
Answers 383 questionsAlignment-as-a-Service: Scale AI vs. the new guys
Answers 383 questionsStop "reinventing" everything to "solve" alignment
Answers 383 questionsOn the current definitions of open-source AI and the state of the data commons
Answers 383 questions
