Compsci Weekly on Nostr: [R] Adam Optimizer Causes Privileged Basis in Transformer Language Models ...

[R] Adam Optimizer Causes Privileged Basis in Transformer Language Models

https://www.lesswrong.com/posts/yrhu6MeFddnGRSLtQ/adam-optimizer-causes-privileged-basis-in-transformer

Discussions: https://discu.eu/q/https://www.lesswrong.com/posts/yrhu6MeFddnGRSLtQ/adam-optimizer-causes-privileged-basis-in-transformer

#compsci #machinelearning