LESSWRONG
LW

eliotpbrenner
0010
Message
Dialogue
Subscribe

Posts

Sorted by New

Wikitag Contributions

Comments

Sorted by
Newest
No wikitag contributions to display.
Towards Multimodal Interpretability: Learning Sparse Interpretable Features in Vision Transformers
eliotpbrenner3mo10

Great work!  I'm trying to reproduce the results.  Using the parameters mentioned in the blog post and associated repo I have started.a training run (but with batch size even larger, 256 instead of 128).  Based on progress so far it looks like it will take about a week to train over all the batches.  Is this what your found @hugofry ?  If that time sounds off to you I'll post my full parameters and ask you to check.  

Reply
No posts to display.