Our results first confirm inverse scaling behavior seen on prior models trained up to 500 zettaFLOPs. But at 2K zettaFLOPs, it becomes U-shaped. U-scaling has also been shown in prior work, such as BIG-Bench.
AI
-
Inverse Scaling Becomes U-Shaped with Larger Language Models
By
–
New preprint!
— Jason Wei (@_jasonwei) 4 novembre 2022
By evaluating 5x larger language models, inverse scaling can become “U-shaped scaling”, which means that performance increases sharply after decreasing.
https://t.co/bZQndKqlB6
These two tasks here are Third Prize winners from the Inverse Scaling Prize. pic.twitter.com/8d3pu8DDrkNew preprint! By evaluating 5x larger language models, inverse scaling can become “U-shaped scaling”, which means that performance increases sharply after decreasing. https://
arxiv.org/abs/2211.02011 These two tasks here are Third Prize winners from the Inverse Scaling Prize. -
JMLR Review Delays: How TMLR Improves Academic Publishing
By
–
I had the same experience with JMLR back in 2015. After 6 months, they hadn't found reviewers. Withdrew the paper and submitted to a different journal. I like that @TmlrOrg doesn't run into this type of friction, since reviewers are expected to review it after being assigned.
-
Improve People Skills for Engineering and Industry 4.0 Leaders
By
–
10 Simple Ways to Improve Your People Skills https://
entrepreneur.com/growing-a-busi
ness/10-simple-ways-to-improve-your-people-skills/343141
… @ASMEdotorg @3DSNorthAmerica @Dassault3DS @cyngn @MargaretSiegien @Cindybolt61 @fogoros @DrFerdowsi @CRudinschi @PawlowskiMario @IIoT_World @MEngineeringMag #Engineering #Industry40 #Robots #AI #Manufacturing -
Democracy increases moral universalism globally survey finds
By
–
Giant worldwide survey looks at universalism (a willingness to treat strangers the same way you treat friends/community members) & finds 26% of people are strongly universalist & 18% are the opposite. Exposure to democracy causally increases universalism. https://
cepr.org/voxeu/columns/
moral-universalism-global-evidence
… -
AI Hardware Enables Real-Time Modern Art Generation at MoMA
By
–
We’re excited to share that our hardware and software will make real-time generation of modern art possible with @RefikAnadol and his upcoming #AIArt exhibit at the @MuseumModernArt
. -
Meta AI Creates Largest Protein Language Model with 15B Parameters
By
–
Meta AI researchers trained a language model to fill in protein sequence gaps across millions of diverse proteins & scaled up to 15B parameters, creating the largest language model of proteins to date. More on our latest breakthrough in protein folding https://
bit.ly/3WoWcK2 -
Meta CEO positioned as winner amid Twitter TikTok outages
By
–
Mark Zuckerberg right now…
— David Herrmann (@herrmanndigital) 4 novembre 2022
Twitter in chaos.
TikTok down. pic.twitter.com/nIJBigO68yMark Zuckerberg right now… Twitter in chaos.
TikTok down. -
Replicate Increases Default Rate Limits for API Predictions
By
–
We've also increased our default rate limits. You can create 10 predictions a second, bursting up to 600 predictions a second. https://
replicate.com/docs/reference
/http#rate-limits
… We can support higher rates too – just email us: team@replicate.com -
Infrastructure improvements deliver faster response times for users
By
–
Over the past few weeks, we've made some major improvements to our infrastructure to make it more reliable and perform better. Nothing's changed from your point of view, but you'll be seeing faster response times!