iEmoji Feeds @Aran Komatsuzaki (arankomatsuzaki)
1.
Aran Komatsuzaki
@arankomatsuzaki
Enhancing GPT-3 with world knowledge:
Introducing REPLUG: a retrieval-augmented LM framework that combines a froz… twitter.com/i/web/status/1…
Introducing REPLUG: a retrieval-augmented LM framework that combines a froz… twitter.com/i/web/status/1…
Retweet of status by @WeijiaShi2
2.
Aran Komatsuzaki
@arankomatsuzaki
Attention the IR community! The era of cheap UNSUPERVISED domain adaptation has begun! Let me introduce the InPars… twitter.com/i/web/status/1…
3.
4.
5.
6.
7.
Aran Komatsuzaki
@arankomatsuzaki
@savvyRL It was nice talking with you
Aran Komatsuzaki
@arankomatsuzaki
Nooo I forgot to include my selfies into LAION dataset, so the models trained on LAION won't know me!
I wonder i… twitter.com/i/web/status/1…
I wonder i… twitter.com/i/web/status/1…
Aran Komatsuzaki
@arankomatsuzaki
Does GPT-3 know arithmetic? Are LM scratchpads/chain-of-thought prompting always helpful? What should go into a s… twitter.com/i/web/status/1…
Retweet of status by @akyurekekin
Aran Komatsuzaki
@arankomatsuzaki
Excited to share RankGen, a 1.2B param contrastive encoder mapping prefixes & model generations to vectors.
larg… twitter.com/i/web/status/1…
larg… twitter.com/i/web/status/1…
Retweet of status by @kalpeshk2011
Aran Komatsuzaki
@arankomatsuzaki
I'm so thankful that 15k people are following me
Now that I have voice, let me talk about my pretty much overlo… twitter.com/i/web/status/1…
Now that I have voice, let me talk about my pretty much overlo… twitter.com/i/web/status/1…
8.
Aran Komatsuzaki
@arankomatsuzaki
Unified Scaling Laws for Routed Language Models
Big scaling paper by Deepmind
arxiv.org/abs/2202.01169
Big scaling paper by Deepmind
arxiv.org/abs/2202.01169
...but wait! There's more!
1.
fakhright
@fakhright
astaghfirullah peng.krim guaaaaaaaa..............a *salto sambil solat*
