Compressing for AGI
Jack Rae
Last updated
Jack Rae
Last updated
相关解读:
原视频:
Think deeply about the training objective of foundation models
What are we doing , why dose it make sense, and what are the limitations?
seek the minimum description length to solve perception
Generative models are lossless compressors
Large language models are state-of-the-art lossless text compressors(?!)
Current limitation of thr apporach
We want to deepest understanding of our observations
ones that generalize