Absolutely brutal. Plenty of us have been warning this is coming with how we're building AI. Need to take a hard look in the mirror and think about if what you're doing is making occurrences like this more or less likely.

Does ChatGPT have a poetic style? Can it generate poems in different forms and styles? In this new paper, we show that, yes, ChatGPT definitely has a poetic style (it’s bad), and it can sort of kind of emulate different styles... except for when it can’t arxiv.org/abs/2410.15299

Radar charts are *very rarely* the best way to visualise information. They're a new research fad, and I want to kill it ☠️. What am I supposed to take away from the below? Too much info for one plot, too hard to disentangle different models, too easy to tell a misleading story 🧵

📢 It's possible to apply for a PhD in data attribution in LLMs, hosted by me, through ELLIS Apply through ELLIS by Nov 15, and also through the formal IT-Universitetet process by Nov 22 (see application links on my website)

We've released the CFP for #FAccT2025, which will be held in Athens, Greece! 🇬🇷 Abstracts are due on January 15th, papers on January 22nd. Read all about it on our new blog: facct-blog.github.io/2024-10-28/ann…

📜Excited to share our comprehensive survey on cultural awareness in #LLMs! 🗺️ We reviewed 300+ papers across diverse modalities (language, vision-language, etc.) Siddhesh Pawar Junyeong Park @ NAACL 2025✈️ Jiho Jin Arnav Arora Junho Myung Inhwa Alice Oh #NLProc openreview.net/forum?id=3gg6G…

Today, Eliot Jones Ivan Yamshchikov Alexander Doria and I are releasing a preprint of our paper, Toxicity of the Commons, as well as the model and dataset we developed. 🧵

I cannot put a finger on it on why now, but 🦋 is bumping with tech folks. And my feed there is from the people I follow. Meanwhile, my feed here is maybe 10% of people I follow, and 90% mainstream media pushed into my face (and too much politics shown that I did not ask for)

I voted! All the way from Denmark! Even though my state is "safe"! Get out there 🇺🇸🇺🇸🇺🇸

The only piece of good news this morning is that in the bluer place, Maria Antoniak made a starter pack of #NLProc people already there. It's pinned in her profile. Feels surreal to have real engagement on posts and see real conversations.

Need to be online, but also hating being online? Consider joining us on Bluesky. It's a bit better there.

Also see Maria Antoniak 's "starter pack" bsky.app/starter-pack/m…

Join the #NLProc party in the other place if you haven't yet! I've never seen a network to grow so fast, at least two hundred new people in the last couple of days. Maria's starter pack has also grown.

I’m thrilled to be at EMNLP this week presenting our paper, “The Empirical Variability of Narrative Perceptions of Social Media Texts” I’ll be giving an oral presentation during the CSS + Cultural Analytics Session 2 (Nov 14). Paper: aclanthology.org/2024.emnlp-mai… 🧵(1/12)

If you aren't on the other place yet🦋, why not?

Nishant Balepur Starter packs make it easier! Maria Antoniak is maintaining a list of NLP researchers on there ✨