FAQ on NeurIPS Europe: NeurIPS Europe is an official NeurIPS 2026 satellite event taking place in Paris, France, alongside the main conference in Sydney and the other satellite event in Atlanta. NeurIPS authors can present their papers at any of the three locations, subject to space availability.

You should now see less unwanted adult content in your For You feed.

Anyone know the fastest way to get new glasses in Seattle? Lenscrafters?

I still always think about the early AI paper (circa 2023?) paper where they get people to guess whether a poem is by a famous poet or ChatGPT and the people perform worse than chance and when you dig a bit deeper into the data you find it’s because they simply hate the real poems

Today is the last day for early registration for #FAccT2026! Register now to join us in beautiful Montréal this summer. Curious what to expect at the conference? Our list of keynote speakers is posted here: facctconference.org/2026/keynote...

“her hobbies included meteorology, music, singing, and video games like Hollow Knight”

New in Nature: LLMs give "the party line" in the languages of authoritarian regimes. This works when they control the media, which feeds pretraining data. We show more state control over the media means less critical LLMs. 6 studies spanning 38 languages & 13 models. Details ↓

Want to be a beta partner of our upcoming product for agentic NLP? We're looking for teams who need help with their NLP projects, especially around tasks like information extraction. Ellf is a virtual assistant for NLP projects – think Claude Code, but with NLP superpowers, or "Vibe NLP" ✨

Such a cool dataset about fiction published in PLAYBOY magazine! "While the magazine is often remembered for its centerfolds and erotica, it also served as a crucial venue for literary work, featuring stories by many of the twentieth century’s most acclaimed and widely read authors."

I'm looking for 1-2 emergency reviewers for a #COLM2026 paper on LLMs and science. Paper looks like a nice read to me! Reply or message me if you might be able to write a review by end of week.

1/ New @Nature! We study how powerful institutions shape the information environment for LLMs. Commercial LLM training is opaque, so we trace a path from state-coordinated media -> training data -> model responses.

Hi everyone, if you speak Portuguese and want to talk about Gemma models, I'll be on this discord channel for the next hours doing a AMA about Gemma! come chat!! discord.com/invite/googl...

“The protocol is open. The tools are available. Blacksky is right there as the working example. Stop trying to renovate someone else's house. Build your own, and let your friends visit, wherever they happen to live.” YES!

Reposting this every time it saves my butt. Maybe it should be built into everything. When you're on the highway, your car should clear its throat now and then: "by the way, do you have 10 min to discuss how my brakes work?" Anyway, thank god PyMC can do that now.

Tired: worrying about catching hantavirus after reading something on the internets whipped up by grifters and alarmists. Wired: worrying about the destruction of infectious disease and pandemic preparedness research and programming by Russell Vought and Donald Trump.

📜 Some of you may have heard about the desk-rejected papers from the ACL'26/ARR Jan 2026 or October 2025 cycles because of hallucinated references. These cases were detected post-commitment. ACL Program chairs have an official statement on this: 2026.aclweb.org/acl_statement/ #ACL2026NLP #NLProc

Submitting to ARR for #EMNLP2026? We're running an opt-in AI Reviewing Experiment. Help us test AI-generated reviews during your ARR submission. 🤖 ✅ Reviewers, ACs, and SACs will not be able to see it ✅ Will not affect decisions 🔗 Read more: 2026.emnlp.org/ai-reviewing...

The project, conducted with Christy Bailey-Tomecek, highlights that while Whisper is efficient, it often 'normalizes' speech. For historical archives, these AI corrections can strip away the nuance of heritage language and personal testimony.

Someone should write something more about this because this myth that people are just using models for info retrieval and code generation keeps coming back, fueled by our own discomfort and company reports that omit or downplay erotic usages of chatbots.

My book 'Artificial Humanities' was reviewed in AI & Society! The review calls the book “the most sustained and theoretically rigorous account to date” of how humanistic knowledge shapes AI, and describes it as a “timely and foundational work.” link.springer.com/article/10.1...

Computational approaches to media narrative analysis either miss nuanced storytelling patterns through coarse-grained analysis, or require domain-specific taxonomies that limit scalability. We show joint event and character modeling can address this gap. Details in our #ACL2026 (Main) paper. 🧵1/10

TIL Anthropic API credits silently expire after one year. What a scammy and unpleasant surprise.

Tomorrow's the day! Our workshop at #LREC2026 on "Leveraging Derived Text Formats to Unlock Copyrighted Collections for Open Science" takes place! Even if you can't make it to the event, you can see the programme and read the papers in the proceedings volume […] [Original post on fedihum.org]

How dare you exploit academic labor; that’s our job!

Legitimately feels like an unquantifiable vibe shift the last few weeks where the pendulum is swinging back to reasonable takes and people experimenting with model choice 🙏

A Human-Centric Framework for Data Attribution in LLMs (FaCCT'26) TLDR: LLMs screwed up data economy, and NLP should help to re-design incentives with data attribution. Here's a moonshot for an alternative LLM use paradigm, for the users, creators & intermediaries. arxiv.org/abs/2602.10995 /1

Also in here: - named entity research experiments - dnd scene descriptions - erotica in the style of Anais Nin - marketing emails - lecture transcripts to be transformed into student notes

I'm annotating some prompts from the newest version of Wildchat, and there are sooo many "Generate a script about Celebrity X or Event Y" that appear (to me) to be intended for "informative" and "interesting" videos shared on social media. I hate it.

What’s the best way to analyze online discourse on any given topic? Is there a right way to use NLP tools to sift through massive datasets? To find out, we tested several tools across different collaboration settings and report findings in an #ACL2026 (Main) paper: arxiv.org/abs/2408.09030 🧵1/7

why "ai for social impact/good" (however you want to call it) should get better at engaging with organizations and institutions that deliver social impact

Chesterton on eugenics, quoted in this great essay on correlation, data visualization, and dual use. Perhaps reminds one of current “il faut” discourse from some AI builders and effective altruists. “it seems clear that they do not mean that the democracy must do it, whatever else they may mean”

"We found that high-quality constitutional documents combined with fictional stories portraying an aligned AI can reduce agentic misalignment" www.anthropic.com/research/tea... Who would have thought to use stories to align LMs? Oh, it was me in 2019... 1/

Bhattacharya goes "there were political machinations in the old process too. (pause) I shouldn't say 'too,' I don't believe there are political machinations in the current process" 🫠🫠🫠

The author sent me a copy, and it's really something to behold in person.

I see lots of posts about the automated reviewing at AAAI. Just to be contrarian: here's a counter from @dirkhovy.bsky.social's team: openreview.net/forum?id=cJh...

Between the Canvas outage and the court win in the NEH grant case, it really feels like a day for @literaturegeek.bsky.social print of @megsmith.bsky.social quote. #DHmakes

the age of the philosopher-coder has arrived. unfortunately it might only be 6 months long

This sounds very cool, much needed perspective on AI! from: > …both a researcher studying language/computation and a reader/writer/citizen of the internet who cares deeply about human creativity, meaning making, and community. Excited to subscribe, thanks for trying on Leaflet @mariaa.bsky.social!

Goodfire released a megapost of all the random feature geometry stuff they're finding, and it's worth a read

The first deadline for submitting a paper to our workshop is today, but there are lots of opportunities to participate: submit a trivia playing robot, play yourself, or write (multimodal) questions. Please share the word! (More details/links in the first comment.)

This database is like if Goodreads existed between 1450 and 1945. "I have read the whole of Jane Austen and think of beginning over again. What a perfect woman" www.open.ac.uk/Arts/reading...

📢 ARR CFP Update Authors with concurrent submissions must discuss them in related work and share anonymized versions. Check the "Originality requirement and overlap with other work by the same authors" section before the ARR May deadline. aclrollingreview.org/cfp #EMNLP #ACL #ARR

“Even if you're well-intentioned and want to comply with this law, it's nearly impossible to do so,” EFF’s Rin Alajaji told @kuer.org. “Our worry here is that this opens the door to other states and other countries to start creating these unworkable mandates.”

Congress may pass a set of dangerous "kids online safety" bills that would censor online speech and harm press freedom, while doing little to protect children. Use our action center to tell Congress: No online censorship in the guise of child safety!

i knew i needed to put that in quotes, caltrain is still miserable (i am on the caltrain, i hate it here)

AFAIK it's not unreasonable to tell a model "don't hallucinate" and hope for some small reduction in hallucinations. See for example: gist.github.com/yoavg/4e4b48...

Every token is a vector in the residual stream; string them together and you get a trajectory. We measured how sharply this path bends (its curvature) in different layers and asked: Do tangled paths produce flatter, more uncertain next-token distributions? (3/n)

Sam Altman texts Mira Murati November 19, 2023

If you work at a Canvas school, like I do, FYI: www.insidehighered.com/news/tech-in...