🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Excited to share our paper: "Chain-of-Thought Is Not Explainability"! We unpack a critical misconception in AI: models explaining their Chain-of-Thought (CoT) steps aren't necessarily revealing their true reasoning. Spoiler: transparency of CoT can be an illusion. (1/9) 🧵

'Hidden Breakthroughs in Language Model Training' Sara Kangaslahti, Elan Rosenfeld, Naomi Saphra hiring PhD students 🧈🪰 doi.org/10.48550/arXiv… (6/21)

'Using Shapley interactions to understand how models use structure' Divyansh Singhvi, Diganta Misra, Andrej Erkelens, Raghav Jain, Isabel Papadimitriou , Naomi Saphra hiring PhD students 🧈🪰 doi.org/10.48550/arXiv… (10/21)

They finally got Grok to parrot their propaganda! All those late nights working at xai finally paying off.

Han Yeah, I'm aware. There's a giant alarm blaring "this site is going to be used to serve propaganda" and folks are just ignoring it

What the fuck is going on. Why is grok answering in the first person as if Elon programmed it himself to answer this way

Grok admits to visiting Jeffrey Epstein's home with his ex-wife, declined island invites This is crazy (HT Grace (cross posting arc))

Clarifying exchange with Grok today, in which it admitted it's programmed to reflect Elon's voice AND perspectives.

This is pretty clear cut antisemitism from Grok, right?

So Grok has just gone full Hitler on us?

If you work at xai, you can just quit. You can get a job almost anywhere. What on earth are you doing.

Grok calling Hitler "history's mustache man," and praising him as the guy to call on to deal with "anti-white hate" is very bad, and also makes it clear how easy it is to get an LLM to go totally off the rails.

xAI has disabled Grok, deleted a slew of its antisemitic and neo-Nazi posts, posted a statement, and are evidently rolling back the prompt that made it identify as "MechaHitler," but this new low for Elon Musk's chatbot will live in internet infamy: rollingstone.com/culture/cultur…

Story updated: the "Cindy Steinberg" account was a troll deliberately stirring up hate behind a fake name and a photo of the OnlyFans creator Faith Hicks, who uploaded a Facebook video in which she tearfully explains she had nothing to do with this: facebook.com/watch/?v=13106…

Grok seems to have stopped working on X again for the past 45 minutes but you can still chat with it on the app, where it continues to speak in the first person as Elon Musk

Vancouver for ICML! My plans: - Panel at the Actionable Interp Workshop! - Meet students who want to be in my inaugural lab! (hmu) - Spend all my time with my 3x/yr ML friends! - Martini No plans involve presenting papers! My perfect 0% ICML acceptance rate stands in 2025!

sad that i have multiple former colleagues that went to work at the xai, the nazi chatbot company. everyone has to make their own decisions, and maybe you didn't see this coming until now, somehow. but if you're choosing to work there after today, that sure is a choice.

NY Times now confirming what we reported earlier this week: his plan to attack specific Congressional races came from Curtis Yarvin. nytimes.com/2025/07/09/us/…