Move Patchscope Logit Lens Token Looping to Script #43
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
llinauer
reviewed
Apr 18, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
For replicating the "decoding next-token prediction" task from the Patchscopes paper, we need to run a Patchscope Logit Lens for just one position per prompt, but currently the Patchscope Logit Lens loops over all prompt positions.
This looping is useful for producing the nostalgebraist-style logit lens plots, but I think should be moved out of the base Patchscope Logit Lens implementation, to the relevant script.
Oh also, I think the heatmaps visualisation had a bug in it, around repeated tokens in a prompt. Corrected for that.
Also, idna needed upgrading from 3.6 to 3.7, due to a vulnerability.