I am disappointed in the AI discourse steveklabnik.com/writing/i-am...

Meta introduced Llama 4 models and added this section near the very bottom of the announcement 😬 “[LLMs] historically have leaned left when it comes to debated political and social topics.” ai.meta.com/blog/llama-4...

"We train our LLMs on art and literature and educational materials, and for some reason they keep turning out progressive."

Have work on the actionable impact of interpretability findings? Consider submitting to our Actionable Interpretability workshop at ICML! See below for more info. Website: actionable-interpretability.github.io Deadline: May 9

Mor Geva@megamor2.bsky.social · last yr.

🎉 Our Actionable Interpretability workshop has been accepted to #ICML2025! 🎉 > Follow @actinterp.bsky.social > Website actionable-interpretability.github.io @talhaklay.bsky.social @anja.re @mariusmosbach.bsky.social @sarah-nlp.bsky.social @iftenney.bsky.social Paper submission deadline: May 9th!

𝐇𝐨𝐰 𝐜𝐚𝐧 𝐰𝐞 𝐩𝐞𝐫𝐟𝐞𝐜𝐭𝐥𝐲 𝐞𝐫𝐚𝐬𝐞 𝐜𝐨𝐧𝐜𝐞𝐩𝐭𝐬 𝐟𝐫𝐨𝐦 𝐋𝐋𝐌𝐬? Our method, Perfect Erasure Functions (PEF), erases concepts perfectly from LLM representations. We analytically derive PEF w/o parameter estimation. PEFs achieve pareto optimal erasure-utility tradeoff backed w/ theoretical guarantees. #AISTATS2025 🧵

New paper from our team @GoogleDeepMind! 🚨 We've put LLMs to the test as writing co-pilots – how good are they really at helping us write? LLMs are increasingly used for open-ended tasks like writing assistance, but how do we assess their effectiveness? 🤔 arxiv.org/pdf/2503.19711

pre aca you would specifically avoid being diagnosed or seeking treatment if you didn't have health insurance to prevent it from making it impossible for you to get health insurance. when you bought health insurance after doing this you committed fraud. i did this.

Some of his readers have asked Mike Masnick @mmasnick.bsky.social why his technology news site, Tech Dirt, has been covering politics so intensely lately. www.techdirt.com/2025/03/04/w... I cannot recommend Mike's reply enough. It's exactly what readers need to hear, what journalists need to do.

Neat visualization that came up in the ARBOR project: this shows DeepSeek "thinking" about a question, and color is the probability that, if it exited thinking, it would give the right answer. (Here yellow means correct.)

Come work with me! We are looking to bring on more top talent to our language modeling workstream at @ai2.bsky.social building the open ecosystem. We are hiring: * Research scientists * Senior research engineers * Post docs (Young investigators) * Pre docs job-boards.greenhouse.io/thealleninst...

The Allen Institute for AI

job-boards.greenhouse.io

I interviewed for LLM/ML research scientist/engineering positions last Fall. Over 200 applications, 100 interviews, many rejections & some offers later, I decided to write the process down, along with the resources I used. Links to the process & resources in the following tweets

Obsessed with the work coming out of Finale Doshi-Velez's group; they don't just take the limits of the real world for ML deployment seriously but instead turn it into new algorithmic ideas arxiv.org/abs/2406.08636

Towards Integrating Personal Knowledge into Test-Time Predictions

Machine learning (ML) models can make decisions based on large amounts of data, but they can be missing personal knowledge available to human users about whom predictions are made. For example, a mode...

arxiv.org

The entire archive of CDC datasets can be found here. HUGE shoutout to data archivists- this work is important 👏🙌🏻 archive.org/details/2025...

Can AI really help with literature reviews? 🧐 Meet Ai2 ScholarQA, an experimental solution that allows you to ask questions that require multiple scientific papers to answer. It gives more in-depth and contextual answers with table comparisons and expandable sections 💡 Try it now: scholarqa.allen.ai

It is such a slap in the face to the Indian American community to delay their green cards for decades and then declare that because of that delay their American children aren't citizens.

Is it time for a social media break again? It has not been a great day. 😅

Mimansa Jaiswal@mimansaj.bsky.social · 2y ago

This probably won’t reach many people, but if someone who feels the same way ends up reading this, I hope it helps them realize they’re not alone. Lately, my mood has been heavily influenced by the ‘tech(adjacent) twitter' vibes, & the past few months have been really rough. ⏎

This is pretty cool! Learn more: developer.chrome.com/docs/devtool... (seems to cover CSS and network requests, which might be fun to lay around with)

i was annoyed at having many chrome tabs with PDF papers having uninformative titles, so i created a small chrome extension to fix it. i'm using it for a while now, works well. today i put it on github. enjoy. github.com/yoavg/pdf-ta...

It's ready! 💫 A new blog post in which I list of all the tools and apps I've been using for work, plus all my opinions about them. maria-antoniak.github.io/2024/12/30/o... Featuring @kagi.com, @warp.dev, @paperpile.bsky.social, @are.na, Fantastical, @obsidian.md, Claude, and more.

Maria Antoniak@mariaa.bsky.social · 2y ago

So far the blog post draft is winning the distraction battle. Prepare for a very long and opinionated update about all the new tools and apps I’ve been using for work.

I've always wanted to build things with D3, but the learning curve was too high. At least for the bespoke stuff I wanted to make (not just simple bar charts). I can finally make things like this thanks to Cursor! I just art directed this, and it made everything work beautifully. Even on mobile 🎉

If you like working with a canvas, Muse (museapp.com) currently has a 30% off. 2 major things that make it different than freeform -- ink sticks to sticky notes (so it feels more like writing in the real world), and you can snippet out sections from pdf that link back to source.

Inspired & focused thinking with Muse

Muse is a canvas for thinking that helps you get clarity on things that matter. Think in private or collaborate with others. Available for iPad and Mac.

museapp.com

I don't usually discuss software, PKM, or tools here, but I found a valuable tip today. Not only can you use this to create beautiful video tutorials similar to what Screenstudio creates automatically, but you can also use this feature during live sharing & streaming, making it incredibly useful!

Josh W. Comeau@joshwcomeau.com · 2y ago

🔥 One of my favourite hidden macOS features is the scroll hotkey gesture. By holding ⌘ (Command) and scrolling down, we zoom WAY IN on the cursor’s location. I learned this trick for highlighting stuff in video tutorials, but it comes in handy a lot in my day-to-day life!

This probably won’t reach many people, but if someone who feels the same way ends up reading this, I hope it helps them realize they’re not alone. Lately, my mood has been heavily influenced by the ‘tech(adjacent) twitter' vibes, & the past few months have been really rough. ⏎

The other major Chinese AI lab, DeepSeek, just dropped their own last-minute entry into the 2024 model race: DeepSeek v3 is a HUGE model (685B parameters) which showed up, mostly undocumented, on Hugging Face this morning. My notes so far: https://simonwillison.net/2024/Dec/25/deepseek-v3/

Our new paper "Fooling LLM graders into giving better grades through neural activity guided adversarial prompting" lead expertly by Atsushi Yamamura shows how even in closed models like Gemini, we can add a small adversarial suffix to an essay to get unreasonably high scores arxiv.org/abs/2412.15275

Fooling LLM graders into giving better grades through neural activity guided adversarial prompting

The deployment of artificial intelligence (AI) in critical decision-making and evaluation processes raises concerns about inherent biases that malicious actors could exploit to distort decision outcom...

arxiv.org

A question for those working on pre-training models (specifically regarding Olmo's logs): at what training step can you begin comparing two models' evaluation performance to get an estimated sense of their relative quality (for example, Olmo7B versus Olmo-2-7B)?

Echoing Theertha Suresh We are hiring! Our team at Google Research NY is seeking a Research Scientist! Our recent research efforts include developing algorithms for improving inference efficiency and alignment of LLMs. If you are interested, please consider applying! www.google.com/about/career...

ImageFX is awesome at creating character sheets, and here are some examples. The text is gibberish though compared to Recraft. I like to use these to add consistent, interesting visuals to presentations with pretty minimal effort. (Prompt expansion by GPT4o in alt)

feeling a but under the weather this week … thus an increased level of activity on social media and blog: kyunghyuncho.me/i-sensed-anx...