-
Notifications
You must be signed in to change notification settings - Fork 427
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Affected files: Monthly Notes/Jun 2024 notes.md
- Loading branch information
swyx
committed
Jun 10, 2024
1 parent
bb0e3ad
commit 9ccbc01
Showing
1 changed file
with
15 additions
and
0 deletions.
There are no files selected for viewing
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
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,15 @@ | ||
|
||
## notable reads | ||
|
||
- [A Picture is Worth 170 Tokens: How Does GPT-4o Encode Images?](https://www.oranlooney.com/post/gpt-cnn/) | ||
- sigma-gpt https://news.ycombinator.com/item?id=40608413 | ||
- The authors randomly permute (i.e., shuffle) input tokens in training and add two positional encodings to each token: one with the token's position and another with the position of the token to be predicted. Otherwise, the model is a standard autoregressive GPT. The consequences of this seemingly "simple" modification are significant: | ||
- The authors can prompt the trained model with part of a sequence and then decode the missing tokens, all at once, in parallel, regardless of order -- i.e., the model can in-fill in parallel. | ||
- The authors can compute conditional probability densities for every missing token in a sequence, again in parallel, i.e., densities for all missing tokens at once. | ||
- The authors propose a rejection-sampling method for generating in-fill tokens, again in parallel. Their method seems to work well in practice. | ||
|
||
## discussions | ||
|
||
|
||
- forcing AI on to us | ||
- msft recall default https://news.ycombinator.com/item?id=40610435 |