This paper looks interesting and important! LLMs can handle more difficult tasks as they increase in size, but they also become less reliable (e.g. more likely to spout hard to spot nonsense): www.nature.com/articles/s41...#MLSky

Larger and more instructable language models become less reliable - Nature

Scaling up and shaping up large language models increased their tendency to provide sensible yet incorrect answers at difficulty levels humans cannot supervise, highlighting the need for a fundam...

The data ceiling is not discussed as much as it deserves to be in public discourse on AI, IMO. tl;dr - AI is already starting to run out of new data: www.economist.com/schools-brie...#mlsky#neuroai

Hi Bluesky! I recently joined, I think it’s time to introduce myself 👋 I’m Leo, #PhD#worms#ecology#neuroscience#pristionchus#nematodes#compneuro#MLsky#neuroskyence

Structured data from Wikipedia, in json format, and including a list of all the entities referred to in a given article. #MLSky 🤖https://enterprise.wikimedia.com/blog/hugging-face-dataset/

Wikipedia Dataset on Hugging Face: Structured Content for AI/ML

Wikimedia Enterprise releasing Wikipedia dataset on Hugging Face, featuring Structured Contents beta from Snapshot API for AI and machine learning applications

We've got three papers coming to EMNLP! 🤖 #mlsky

Are you looking to do an independent postdoc on a research topic involving data science, machine learning, or AI? Apply to be a faculty fellow at the Center for Data Science at NYU! 2-year position, including some teaching, at a very interdisciplinary center! #MLsky#CCAIapply.interfolio.com/153414

Tak když jdeme na tah, vždycky s náma jde pes a nikdy nám nikdo nic neřekl. Naopak 😁 Ještě se chodí lidi mazlit a nosí vodu a mlsky 😊

I have published a detailed look at ChatGPT's ability to identify patterns in HTR errors . Here is the link: www.linkedin.com/pulse/workin...#machinelearning#MLSky@thiagokrause.bsky.social@anterotesis.bsky.social

Working with ChatGPT oi preview: Error recognition and classification in raw machine transcriptions of C17th legal depositions

Overview MarineLives is running a project to machine transcribe and machine/human correct 50,000 pages of legal depositions from the C17th. That's roughly 30 million words.

Mlsky koupené za stravenky nejenže jsou zadarmo, ale mají navíc nulovou kalorickou hodnotu.

We’ve recently put together a list of potential Journal Club papers for our lab. If curious what papers we may cover in a journal club, voilà: snailab.notion.site/c5ec9ee7bbf7...#neuroscience#MLSky 🧠🤖

Notion – The all-in-one workspace for your notes, tasks, wikis, and databases.

A new tool that blends your everyday work apps into one. It's the all-in-one workspace for you and your team