The same can probably be said also about a lot of the compliance and risk functions in major companies.

There are just too many options out there now. I think most people got desensitized. It's just another LLM in the end.

Yeah, there isn't that novelty anymore compared to the deepseek moment. Maybe even expected by many in the Chinese tech industry.

I thought most papers of this size just use the name of the consortium, like the human genome project did with many of its papers

Do you have the link at hand? :)

I guess asking trusted people is the real definition of a peer review then

He wanted to test how models performed when subjected to a new tokenization scheme without re-training the model.
Modifying Llama 3’s tokenizer to tokenize numbers from right to left (R2L) instead of left-to-right (L2R) with just a few lines of code 🧑‍💻. This affect the grouping of numbers by three:

It's just so freaking fast!

Wish it was also available in Python