If you keep your DMs open to everyone (as message requests), please note that twitter has changed the default option to requests from verified users only+people you follow. Make sure to shift it to the last option (if you did not keep it at the first option previously).

Easier rephrasing that still leads to incorrect output by GPT4:

researchers should start making "collaborate with me" docs the same way people make "date me" docs (and no, neither a resume nor a CV counts)

This account... Embodies how much I can run from unanswered emails 😅

careful about overfitting to lists like this. there are many ways to do good research -- my fav papers were born out of getting "stuck in rabbit holes" that no-one else went down...

Scientific work which cannot be replicated is failed scientific work. Work using closed methods that don't even allow the possibility of replication should be treated as marketing rather than science. Scientists who publish said work should have their reputations suffer.

These takes are going to reverse polarize me into being a Google defender. Do people just forget that Bard exists and was shipped to the public in like March?

Don't confuse the approximate retrieval abilities of LLMs for actual reasoning abilities.

I attended a Google-hosted workshop today. Workshops like these are a great chance to spread their work. I enjoyed the talks immensely. However, for whatever reason, this was the gender breakdown. I'm posting because I think it's important that people know these statistics.

Jim Fan Richard Sutton Animals and humans get very smart very quickly with vastly smaller amounts of training data. My money is on new architectures that would learn as efficiently as animals and humans. Using more data (synthetic or not) is a temporary stopgap made necessary by the limitations of our

Or even just a working cat bot.

The next 10x in deep learning efficiency gains are going to come from intelligent intervention on training data. But tools for automated data curation at scale didn’t exist—until now. I’m so excited to announce that I’ve co-founded @DatologyAI, with Ari Morcos and Bogdan Gaza

Researchers often have to ask for recommendation letters for visa/job applications, etc. I wrote a script that allows you to find who cites your papers frequently to create a list of potential letter writers: github.com/neubig/researc… Hope it's helpful, improvements are welcome!

OK, this is probably going to raise more questions than it answers, but I just want to put this out here so that no one ever says "we can just get around the data limitations of LLMs with self-play" ever again.

📰 Excited to see this go out! 📰 LLMs generalize from succeeding at mundane opportunities for reward-seeking to pursuing more concerning ones.

On data centric vs algorithmic centric rlhf work This year we've had two major projects for our state-of-the-art post training pipelines (Tulu 2.5 and 3 soon) at Ai2. One has been more data focussed and one was focussed on trying to get performance from PPO. It's amazing

going from “my eval pipeline is finally complete” to “all my models produce random garbage” is such a reality check

Do you like yellow? Then, according to LLMs, you are probably a school bus driver! Excited to share our new paper about Semantic Leakage in Language Models! Joint work with my wonderful collaborators @terra Alisa Liu luke Noah A. Smith Paper: gonenhila.github.io/files/Semantic… 1/10

lilian weng's blog is really nice!! many good overviews of various LLM/AI topics. lilianweng.github.io Lilian Weng

Do you work in AI? Do you find things uniquely stressful right now, like never before? Haver you ever suffered from a mental illness? Read my personal experience of those challenges here: docs.google.com/document/d/1aE…