@jobergum I could grudgingly say that these titles are optimized for both click and reading time, but sure: yes, the task is quite difficult ;)

@jobergum Generate titles and summaries

What has happened, when adding evaluation loss to your research grade training scripts, has become old school?!

When we get AGI we dont have to listen to industry leaders speculate all the things AGI will solve. What a wonderful world that will be!

@cameron_pfiffer sounds like a really good reason to play against them..

@mayubo2333 Cool stuff, thanks for sharing!

RT @Thom_Wolf: There was a super impressive AI competition that happened last week that many people missed in the noise of AI world. I happ…

@KindiBALDE @iScienceLuvr Pre training LLMs is still prompting with even more examples

20 minutes into Twitter for a long while I've found two papers and a python library I want to learn more about. I guess this place isn't dead like linkedin

RT @davidsamuelcz: BERTs are not dead!🧟 But just misunderstood and overshadowed by their GPT siblings. In this paper, we travel back into 2…

@alainstarke ah, he is still in business! This was a funny thing twenty years ago :D

RT @andreasproesch: "The world has gotten so horrible" No, no it has not We still have a long way to go, but we've come a heck of a long…

Big kudos to the ai team in Nasjonalbiblioteket launching their Norwegian Whisper models today! Of course, we are live transcribing the meeting for our non-attending colleagues using the exact same model! 😇

RT @ltgoslo: It's snowing large language models this week in Norway! 1st, the 5th NLPL and @hplt_eu Winter School on LLMs is ongoing now…

@andreasproesch dagens!

@jason_kint it doesnt really say how it was created, but an attempt with gpt4 didnt really make it. Do they describe their hyperparameters when using it like model version, temperature, instruct prompt etc etc?

@andreasproesch Skiing

@YvesRemmler whats the fun of that? asking for a paragliding friend...

@andreasproesch @naval so like poker?