←
Rob K2,610 posts
Rob K avatar

Rob K

@robertzzk

Focused on AI work.

◷ Joined October 2014▧ 301 Photos and videos
Rob K avatar
⟳ Rob K retweeted
Rob K@robertzzk·7 Dec 2025

TransformerLens 3.0 is out! 🎉 Congrats to collaborator and friend Bryce Meyer (no X 😢) for getting it across the line. Details in the Open Source Mechanistic Interpretability slack: opensourcemechanistic.slack.…

Spencer Cheng avatar
⟳ Rob K retweeted
Spencer Cheng@spenccheng·11 Jul 2025

Joseph's guide provides clear tactical advice on how to learn RL. Give it a read. Build Environments. You can just learn RL.

Joseph Suarez 🐡 @jsuarez 11 Jul 2025 Article An Ultra Opinionated Guide to Reinforcement Learning Reinforcement learning is about learning through interaction. Applications include robotics, logistics, gaming, and even control problems in science like nuclear fusion. It's an underexplored niche of

Iván Arcuschin avatar
⟳ Rob K retweeted
Iván Arcuschin@IvanArcus·13 Mar 2025

🧵Chain-of-Thought reasoning in LLMs like Claude 3.7 and R1 is behind many recent breakthroughs. But does the CoT always explain how models answer? No! For the first time, we show that models can be unfaithful on normal prompts, rather than contrived prompts designed to trick them.

Post attachment
Peter Wildeford🇺🇸🚀 avatar
⟳ Rob K retweeted
Peter Wildeford🇺🇸🚀@peterwildeford·1 Oct 2024

🚨 IAPS is hiring 🚨 We seek Researchers / Senior Researchers to join our team to identifies concrete interventions that could manage risk and secure benefits of advanced AI. One role is fully remote and can be done in most countries, the other will be in DC. App due Oct 21

Post attachment
Rob K avatar
Rob K@robertzzk·19 Jul 2024

Outer loss transmits very limited bits of information through the info-theoretic bottleneck of the training process... so yes, if your function family (eg wetware brains) priors mesa-optimizers relatively densely, no sparsely delivered outer loss (death in the EEA) suffices...

"Train it to be nice" is the obvious thought. Alas, I predict that one idiom that does generalize from natural selection to gradient descent, is that training on an outer loss gets you something not internally aligned to that outer loss. It gets you ice cream and condoms.

Rob K avatar
Rob K@robertzzk·19 Jul 2024

Disagree. "Gradient-descenting" is the wrong verb as it implies attribution to kindness-likelihood is primarily the optimizer. The text that is to be predicted matters; the etiology of the ontology implicit in the training distribution *was* shaped by m-s of proteins and IS kind

Gradient-descending an AI system to predict text, or even to play video games, is nothing like this. It is exploring nowhere near this space. Gradient descent of matrices is not mutation-selection of proteins and I don't expect it to hit on anything like similar architectures.

Akshay Kothari avatar
⟳ Rob K retweeted
Akshay Kothari@akothari·19 Jul 2024

The Microsoft / CrowdStrike outage has taken down most airports in India. I got my first hand-written boarding pass today 😅

Post attachment
Rob K avatar
Rob K@robertzzk·12 Jul 2024

Open X for *one* moment and then it's, like, 3 hours of ML paper reading later... 👋 cya crazy peeps

Rob K avatar
Rob K@robertzzk·12 Jul 2024

My feed is nothing except COLM paper acceptances. I get, it you got in, nice job. I'm very proud of you.

Rob K avatar
Rob K@robertzzk·27 Jun 2024

Sparse Autoencoders help us understand the MLPs of LLMs, but what's up with attention? Find out in our new paper with @Connor_Kissane and @NeelNanda5!

Sparse Autoencoders help us understand the MLPs of LLMs, but what's up with attention? In our new paper with @NeelNanda5, we introduce Attention Output SAEs to uncover what concepts attention layers learn. Further, we use them to find novel insights previously out-of-reach!🧵

Post attachment
Rob K avatar
Rob K@robertzzk·22 Jun 2024

New attention sparseautoencoders post dropped. $1000 bounty to whomever find the best attention circuit!

Neel Nanda@NeelNanda521 Jun 2024

Great post from my scholars @Connor_Kissane & @robertzzk! SAEs are fashionable, but are they a useful tool for researchers? They are! We find a deeper understanding of the well-studied IOI circuit, and make a circuit analysis tool $1000 bounty to whoever finds the best circuit!