We're recruiting a student researcher for autumn 🍁 to work with Sunipa Dev, Remi Denton and myself on LLMs, fairness, safety, and inclusion. #NLProc ↘️

We're hiring a data engineer to help with Dolma! Come work with me on:
- data acquisition 🕷️
- high–performance data pipelines ⚡️
- open source science 🔄

Please apply if you have experience in any of those! DM for questions, job link in thread 🧵

“Can we get a new text analysis tool?”
“No—we have Topic Model at home”

Topic Model at home: outputs vague keywords; needs constant parameter fiddling🫠

Is there a better way? We introduce LLooM, a concept induction tool to explore text data in terms of interpretable concepts🧵

'15T tokens collected from publicly available sources'.

what does 'publicly available source' even mean?

An unintended consequence of letting GPT4 write all the reviews? Just saying, but if I were a system that had achieved GAI, the first thing I would do is spike all the LLM understanding research

I already use “high school pitcher” from Moneyball as a shorthand for PhD applicants who in it for the wrong reasons and are likely to flame out, I guess now we’ll be able to identify them faster

By the way, we already have three statutory definitions of AI (for those of us keeping count).

Pic 1. Section 238(g) of the John S. McCain National Defense Authorization Act for Fiscal Year 2019 (P.L. 117-232). (uscode.house.gov/view.xhtml?req…)

Pic 2. The National Artificial…

I love doing this kind of hand-tooled programming, but PL folks, isn’t this supposed to be what compilers are for?

This should be part of a task on common sense implication

ICYMI I'm looking for strong postdoc candidates in the space of human-AI interaction!

Sooooo… tweaks

🚨NLP+CSS workshop is back and will be at NAACL 2024!

Paper submission deadline: March 24 sites.google.com/site/nlpandcss/

Organizing team: Anjalie Field Dallas Card Dirk Hovy and myself