Stars and mentions
Last 30 days, Oct 4 – Nov 2
Package downloads
Top repositories
deepmind’s repositories ranked by new stars, as on trending.
Contributes to
Repositories outside deepmind’s account where deepmind opened or reviewed pull requests in the past 365 days.
Who noticed deepmind
Everyone who starred or forked deepmind’s repositories since the archive began in 2011.
Notable accounts
Ordered by the stars their own repositories have received.
Biggest fans
Accounts that starred the most of these repositories.
Breakout weeks
Each repository’s best week of new stars, with the release published around it.
- Gold
- 91st place finishes
- Silver
- 52nd place finishes
- Bronze
- 93rd place finishes
- Awarded repositories
- 11Owned by deepmind now
Repositories
Ordered by gold, then silver, then bronze.
- alphafoldBest: 1st by stars, 1,956 stars410
- mujocoBest: 1st by stars, 725 stars212
- sonnetBest: 1st by stars, 1,077 stars202
- alphatensorBest: 1st by stars, 621 stars100
- graph_netsBest: 2nd by stars, 576 stars011
- code_contestsBest: 2nd by stars, 320 stars010
- pysc2Best: 2nd by stars, 939 stars010
- labBest: 3rd by stars, 230 stars001
- learning-to-learnBest: 3rd by stars, 628 stars001
- open_spielBest: 3rd by stars, 328 stars001
- trflBest: 3rd by stars, 625 stars001
Every award
Top-three finishes of these repositories in completed trending days and weeks, newest first.
- #1 by starsDaily trending
621 starsalphatensor - #2 by starsDaily trending
320 starscode_contests - #3 by starsDaily trending
96 starsmujoco - #3 by starsDaily trending
145 starsmujoco - #1 by starsDaily trending
274 starsmujoco - #1 by starsWeek of
725 starsmujoco - #2 by starsDaily trending
169 starsmujoco - #1 by starsDaily trending
588 starsalphafold
A repository earns an award for finishing first, second or third in a completed trending day or week, ranked by new stars, new forks or pull request authors. Awards follow a repository through renames and transfers and count toward its current owner. Weeks run Monday to Sunday, UTC.
Mentions
Here's the work that shows how rule-based logic can be compiled into transformer weights: * Thinking Like Transformers” by Weiss, Goldberg & Yahav (ICML 2021).: arxiv.org * Tracr: Compiled Transformers as a Laboratory for Interpretability arxiv.org (code at deepmind/tracr) So it's not really an "LLMs can modify weights" thing. Rather it's "LLMs can write programs that can write or modify weights". Indeed LLMs can write many programs, it's impressive.