←
tom white4,249 posts
tom white avatar✓

tom white ✓

@dribnet

creations with code and networks

◷ Joined June 2011▧ 1,512 Photos and videos
tom white avatar
tom white@dribnet·Oct 22

To be clear: we urgently need a few sponsors to support a bold instagram friendly artwork installation adjacent to registration as well as the above art exhibition in the sold out expo hall where your own branding can be prominently displayed in appreciation. 🙏🤞

tom white avatar
tom white@dribnet·Oct 21

NeurIPS Creative Track 2026 is seeking sponsors to help fill high-visibility public areas of the NeurIPS convention center with art. If you or your company would like to help make this happen, please DM me this week!

Post attachment
tom white avatar
tom white@dribnet·Jul 8

just some early results - slightly cherry-picked, but things seem to improve with scale which is a good sign. kudos to @wesg52 @sofroniewn @Jack_W_Lindsey and team for the refrence implmentation which made it easy to explore!

Post attachment
5:27
tom white avatar
tom white@dribnet·Jul 8

here's the same effect with a different print: on the 48 layer gemma3-12b, at layer 47 you get bird, woodpecker, ornothology but upstream at layer 37 its again just "bird birds birds"

Post attachmentPost attachmentPost attachment
tom white avatar
tom white@dribnet·Jul 8

previously my main intuition pump nearest neighbors in vision models (shown here). my goal is always to make drawings that generalize to new models and techniques so happy that understanding also works here. amusingly if you go upstream responses simplify to just "books books books".

Post attachmentPost attachment
tom white avatar
tom white@dribnet·Jul 8

enjoying the new global workspace anthropic paper - especially using the Jacobian lens to provide readout interpretations of prints shown at last month CVPR

Post attachmentPost attachment
tom white avatar
tom white@dribnet·Jun 6

If you are at CVPR drop by - six pieces are on display today and tomorrow just behind the Google booth in the #CVPR2026 art show. I'll be there for daily artist talks on the official schedule or DM if you'd like to meet up and chat.

Post attachment
tom white avatar
tom white@dribnet·Jun 6

Each print's unique backlighting element is provided by a TIPS model. This shows spatial patch level activations highlighting visually where that model attends to when activating its analogous feature.

Post attachmentPost attachment
tom white avatar
tom white@dribnet·Jun 6

Behind the scenes I have a basket of metrics to evaluate how the prints are interpreted. For example, in the spirit of Platonic Representation Hypothesis I evaluate mutual knn - using the print itself to gauge concept coherence across models.

Post attachmentPost attachment
tom white avatar
tom white@dribnet·Jun 6

pleased to be showing new screenprints at @CVPRConf this weekend. these use mechanistic interpretability techniques to highlight features shared across a wide family of vision models

Post attachmentPost attachment
tom white avatar
tom white@dribnet·Mar 14

the models, they just wanna converge

tom white avatar
tom white@dribnet·Mar 14

saxophone print + top-150 SigLIP image probe though mknn model agreement (a la platonic representation hypothesis) is not part of the test time compute process, it climbs naturally as the as the optimization evolves

Post attachmentPost attachment
tom white avatar
tom white@dribnet·Mar 12

45% mutual kNN between CLIP and SigLIP — not bad for two model families trained on different data with different objectives when probed with this print. revisiting ImageNet so I can build a toolbox for navigating more uncharted waters without class labels (stay tuned)...

0:06
tom white avatar
tom white@dribnet·Mar 10

unsure how AI interprets this print? treating the image as a linear probe on your favorite vision model and scraping a diverse dataset for maximum activations provides a coherent suggestion.

0:20
tom white avatar
tom white@dribnet·Mar 6

or query your favorite vision model for semantic nearest neighbors - here's OpenAI-CLIP's top hits across CC3M using the baseball_player print as a probe

Post attachment