Nate Rahn
@n8rahn
Research Fellow @AnthropicAI, PhD student @mila_quebec, formerly @Google eng, @BrownUniversity. Making LLMs explorative, adaptive, and goal-oriented
We will present our work as poster #1423 at NeurIPS next Tuesday 12 Dec at 10:45 am. Come chat with us about the empirical science of neural network-based agents! @n8rahn @harwiltz @pierrelux @marcgbellemare @Mila_Quebec
We built a map of the behaviors learned by deep reinforcement learning agents and we found some surprises! In our NeurIPS 2023 paper, “Policy Optimization in a Noisy Neighborhood: On Return Landscapes in Continuous Control”, we study the return landscape, the mapping from an…
Check out our new discoveries on the empirical science of deep reinforcement learning!
We built a map of the behaviors learned by deep reinforcement learning agents and we found some surprises! In our NeurIPS 2023 paper, “Policy Optimization in a Noisy Neighborhood: On Return Landscapes in Continuous Control”, we study the return landscape, the mapping from an…
Late update: I’ve moved to the Bay Area for a 6-month research fellowship at @AnthropicAI ! I’d be glad to meet other researchers working on RL for language models, agents, subtle and unverifiable rewards, etc. — DMs open.
Proud to have been part of the team behind Meta Motivo, a truly groundbreaking foundation model for behavior. It’s the first of its kind, enabling you to instantly generate human-like behaviors for any reward function or goal. Make sure to check out the demo for yourself!
New release from Meta FAIR — Meta Motivo is a first-of-its-kind behavioral foundation model for controlling virtual physics-based humanoid agents for a wide range of complex whole-body tasks. The model is capable of expressing human-like behaviors and achieves performance…
With today's announcement @karlmoritz, Richard & I are thrilled to launch Reliant's next phase - building AI that will completely change how we work with data. Excited to bring Tola Capital, @inovia, and @mavolpi's expertise & experience on this journey. PS: We're hiring :)
Thanks @TechCrunch for covering our $11.3M seed round, bringing next gen(AI) analytics to biopharma and beyond. techcrunch.com/2024/08/20/rel… Happy to have great investors on board with Tola Capital, @inovia and @mavolpi in additon to our amazing Angels from before.
New #RLC2024 paper Three Dogmas of Reinforcement Learning joint w/ @mark_ho_ and @aharutyu! arxiv.org/pdf/2407.10583 We reflect on where our scientific paradigm needs adjustment, and suggest three departures from previous conventions. Curious to hear what folks think! 🧵
Off to #ICML2024 to present our work on “Controlling Large Language Model Agents with Entropic Activation Steering” at the mech interp wkshp. Would love to meet folks curious about understanding/improving LLM agents, steering vectors, etc - DM or email me if you'd like to chat!
Did you miss the recent Auroras? No problem! ✨🎆 Super excited to share AURORA, a *general* image editing model + high-quality data that improves where prev work fails the most: Performing *action or movement* edits, i.e. a kind of world model setup Insights/Details ⬇️
Excited to share the latest paper of my PhD, looking at steering LLM agents. Check out the thread!
WHAT?? You can steer an LLM agent’s representation of uncertainty?? Introducing Entropic Activation Steering (EAST), a method for controlling an LLM agent's uncertainty in its decisions. EAST computes a steering vector by an entropy-weighted average of representations and uses…
Introducing the Distributional Successor Measure (DSM): a model of the range of possible futures an agent faces. As a distributional extension of the Successor Representation, it enables zero-shot distributional policy evaluation beyond the capabilities of existing methods. The…