Excited to share our #ICLR2024 work led by Megan Richards on geographical fairness of vision models! 🌍 We show that even the SOTA vision models have large disparities in accuracy between different geographic regions. openreview.net/pdf?id=rhaQbS3…

As part of our commitment to open science, I'm excited to share that an alpha version of our protein design code is available! Check out the tutorials to learn how to designs proteins yourself with guided discrete diffusion, just like the pros github.com/prescient-desi…

📢Workshop on Reliable and Responsible Foundation Models will happen today (8:50am - 5:00pm). Join us at #ICLR2024 room Halle A 3 for a wonderful lineup of speakers, along with 63 amazing posters and 4 contributed talks! Schedule: iclr-r2fm.github.io/#program.

An image is worth more than one caption! In our #ICML2024 paper “Modeling Caption Diversity in Vision-Language Pretraining” we explicitly bake in that observation in our VLM called Llip and condition the visual representations on the latent context. arxiv.org/abs/2405.00740 🧵1/6

New Anthropic research paper: Scaling Monosemanticity. The first ever detailed look inside a leading large language model. Read the blog post here: anthropic.com/research/mappi…

Please submit your work on robustness / privacy / trustworthiness / alignment / ... with multimodal foundation models to our ICML workshop!

I’m excited to announce that I’ll start as an assistant professor at Columbia University this summer! Interview season was fun, I met so many amazing people, but I’m happy to finally close the loop.

I'm excited to join Anthropic to continue the superalignment mission! My new team will work on scalable oversight, weak-to-strong generalization, and automated alignment research. If you're interested in joining, my dms are open.

Excited to share what I've been working on as part of the former Superalignment team! We introduce a SOTA training stack for SAEs. To demonstrate that our methods scale, we train a 16M latent SAE on GPT-4. Because MSE/L0 is not the final goal, we also introduce new SAE metrics.

Whether LLMs can reliably be used for decision making and benefit society depends on whether they can reliably represent uncertainty over the correctness of their outputs. There's anything but consensus. In new work we find LLMs must be taught to know what they don't know. 1/6

Modern generative models are trained to imitate human experts, but can they actually beat those experts? Our new paper uses imitative chess agents to explore when a model can "transcend" its training distribution and outperform every human it's trained on. arxiv.org/abs/2406.11741

Don't miss our tutorial A Bayesian Odyssey in Uncertainty: from Theoretical Foundations to Real-World Applications European Conference on Computer Vision #ECCV2026 on Mon AM (08:45-13:00) #ECCV2024 In the line-up we have: Pavel Izmailov Gianni Franchi Alexander Immer Olivier Laurent Adrien Lafage uqtutorial.github.io

Excited to give an invited talk tomorrow Sep 30 at the #ECCV workshop on Uncertainty Quantification in Computer Vision at 12:25pm CET in room Brown3! I will present our research on spurious correlations and geographic biases in large-scale vision and multi-modal models!

Apply to join the Anthropic Fellows Program! This is an exceptional opportunity to join AI safety research, collaborating with leading researchers on one of the world's most pressing problems. 👇 alignment.anthropic.com/2024/anthropic…

Watch this special debate live tomorrow at 10:30 a.m. PT — part of our workshop on Unknown Futures of Generalization. Register to attend in person, access the livestream, or view the recording before it's captioned for publication: simons.berkeley.edu/web-registrati…

🚀 Are you passionate about #Alignment Research? Exciting news! Join us at the ICLR 2025 Workshop on 👫<>🤖Bidirectional Human-AI Alignment (April 27 or 28, Singapore). We're inviting researchers in AI, HCI, NLP, Speech, Vision, Social Science, and beyond domains to submit their

Introducing Claude 3.7 Sonnet: our most intelligent model to date. It's a hybrid reasoning model, producing near-instant responses or extended, step-by-step thinking. One model, two ways to think. We’re also releasing an agentic coding tool: Claude Code.

🚀 #ICLR2025 & #CHI2025 are just around the corner — and we’re excited to welcome you to our Bidirectional 👫Human-AI🤖 Alignment events! 🏅 "Golden Sponsors"🏅 A heartfelt thank you to our two generous Golden Sponsors: 🌟Prolific and 🌟Layer 6 AI of TD Bank! Their

Polina is great, and this seems like a great opportunity!