Advertisement · 728 × 90

Posts by Jörg Bornschein

LLMs struggle to transfer knowledge between languages.

Fascinating paper with some intriguing experiments.

arxiv.org/pdf/2502.21228

11 months ago 2 0 0 0

... inspired by Terry Bisson story "They're Made Out of Meat".

www.mit.edu/people/dpoli...

11 months ago 0 0 0 0
Preview
_its_not_real_ on X: ""They're made out of meat." "Meat?" "Meat. Humans. They're made entirely out of meat." "But that's impossible. What about all the tokens they generate? The text? The code?" "They do produce tokens, but the tokens aren't their essence. They're merely outputs. The humans themselves" / X "They're made out of meat." "Meat?" "Meat. Humans. They're made entirely out of meat." "But that's impossible. What about all the tokens they generate? The text? The code?" "They do produce tokens, but the tokens aren't their essence. They're merely outputs. The humans themselves

By x.com/_its_not_rea...
[Apologies, link to X]

11 months ago 0 0 0 0

"They're made out of meat."
"Meat?"
"Meat. Humans. They're made entirely out of meat."
"But that's impossible. What about all the tokens they generate? The text? The code?"
"They do produce tokens, but the tokens aren't their essence. They're merely outputs. The humans themselves are meat."
[...]

11 months ago 0 0 2 0
Post image

Large language models store vast amounts of knowledge, but how exactly do they learn it?

Excited to share my Google DeepMind internship results, which reveal the fascinating dynamics behind factual knowledge acquisition in LLMs!

1 year ago 29 3 1 2
Intro to DeepSeek's open-source week and why it's a big deal Quick Intro to FlashMLA, DeepEP, DeepGEMM, DualPipe, EPPLB, 3FS and Smallpond

Summary of the DeepSeek open source releases last week:
www.pyspur.dev/blog/deepsee...

1 year ago 0 0 0 0
Post image

Really interesting exam of context influence on conceptual mapping in LLMs (also clever) arxiv.org/abs/2501.00070

1 year ago 27 5 0 0
Preview
Accelerated Diffusion Models via Speculative Sampling Speculative sampling is a popular technique for accelerating inference in Large Language Models by generating candidate tokens using a fast draft model and accepting or rejecting them based on the tar...

arxiv.org/abs/2501.05370

1 year ago 1 0 0 0
Preview
Liebreich: Generative AI – The Power and the Glory | BloombergNEF This year will go down in history as the year the energy sector woke up to AI. This is also the year AI woke up to energy. Is the data center power frenzy just the latest of a long line of energy sect...

What happens when techno-optimism meets energy system realism? My latest for @BloombergNEF is a long read on AI, called The Power and the Glory. So that's Boxing Day sorted!
about.bnef.com/blog/liebrei...

1 year ago 160 53 21 22
Advertisement
CALL FOR WORKSHOPS | RLDM

The RLDM workshop list is now up! Have a look: rldm.org/call-for-wor...

More workshop details coming soon :)

There were so many wonderful submissions---this was really tough for the committee. Huge thanks to all involved, and looking forward to seeing folks in Dublin

1 year ago 42 14 2 0

For my first post on Bluesky .. I'll start by announcing our 2025 edition of EEML which will be in Sarajevo :) ! I'm really excited about it and hope to see many of you there. Please follow the website (and Bluesky account) for more details which are coming soon ..

1 year ago 32 7 1 0

Am I the only one struggling to understand how Docker's automatic iptables rules interact with my own when managing nat/open/forwarded ports?

... I struggle to build a reliable mental model of how the automatic and manual rules in all those chains interact.

1 year ago 2 0 0 0