NDIF Team

Making model internals accessible to domain experts in low-code interfaces will unlock the next step in making interpretability useful across a variety of domains. Very excited about the NDIF Workbench! 💡

👀 More advanced interpretability tools coming soon. What techniques would you like to see? Reach out or drop suggestions in the form.

NDIF Workbench Feedback Thank you for taking the time to submit your feedback! Every little bit helps.

This is a public beta, so we expect bugs and actively want your feedback: forms.gle/WsxmZikeLNw3...

YouTube video by NDIF Team Workbench Logit Lens Demo

Study any NDIF-hosted model (including Llama 405B) directly in your browser. Our first tool, Logit Lens, lets you peer inside LLM computations layer-by-layer. Watch the full demo on YouTube (www.youtube.com/watch?v=BK-q...) or try it yourself: workbench.ndif.us

Ever wished you could explore what's happening inside a 405B parameter model without writing any code? Workbench, our AI interpretability interface, is now live for public beta at workbench.ndif.us!

👀 More advanced interpretability tools coming soon. What techniques would you like to see? Reach out or drop suggestions in the form.

NDIF Workbench Feedback Thank you for taking the time to submit your feedback! Every little bit helps.

This is a public beta, so we expect bugs and actively want your feedback: forms.gle/WsxmZikeLNw...

Read the paper or play around with some demos on the project website!

ArXiv: arxiv.org/abs/2410.22366
Project Website: sdxl-unbox.epfl.ch/

In this talk, Chris Wendler presents his recent work on using sparse autoencoders for diffusion models. In this work, they train SAEs on SDXL Turbo, finding ... Interpreting SDXL Turbo Using Sparse Autoencoders with Chris Wendler

New YouTube video posted! @wendlerc.bsky.social presents his work using SAEs for diffusion text-to-image models. The authors find interpretable SAE features and demonstrate how these features can alter generated images.

Watch here: youtu.be/43NnaqGjArA

NSF National Deep Inference Fabric NDIF is a research computing project that enables researchers and students to crack open the mysteries inside large-scale AI systems.

Reminder that today is the deadline to apply for our hot-swapping program! Be the first to test out many new models remotely on NDIF and submit your application today!

More details: ndif.us/hotswap.html
Application link: forms.gle/KHVkYxybmK12...

Want increased remote model availability on NDIF? Interested in studying model checkpoints?

Sign up for the NDIF hot-swapping pilot by October 1st: forms.gle/Cf4WF3xiNzud...

NDIF Hot-swapping Beta Testing Do you have a research project where you plan to study many different models? NDIF will soon be deploying model hot-swapping, which will enable users to access any HuggingFace model remotely via NDIF. We are soliciting applications for a pilot program to beta test our hot-swapping functionality on real research. By participating, you will: Be in the first cohort of users to access models beyond our whitelist (including checkpoints) Directly control which models are hosted on the NDIF backend Receive 1:1 research and technical support from NDIF team Give feedback to NDIF, guiding future user experience Application Information: Apply by October 1st, 2025 Acceptance by October 15th, 2025 Applications will be reviewed based on impact, feasibility, and fit with the NDIF/NNsight platform We are particularly interested in supporting work on model checkpoints and training dynamics Please email info@ndif.us with any questions.

Apply by October 1st, 2025: forms.gle/YxtRh83r5cQ...

For any questions about the application or the NDIF platform, please contact us at info@ndif.us.

Participants will:

1. Be in the first cohort of users to access models beyond our whitelist
2. Directly control which models are hosted on the NDIF backend
3. Receive guided support on their project from the NDIF team
4. Give feedback, guiding future user experience

This fall, we are running a program to test our model hot-swapping on real research projects. Projects should require internal access to multiple models, which could include model checkpoints, different model sizes, unique model architectures, or other creative approaches.

Do you wish you could run experiments on any model remotely from your laptop? In a future release, NDIF users will be able to dynamically deploy any model from HuggingFace on NDIF for remote experimentation. But before this, we need your help!

09:30 AM - 09:40 AM: Opening Remarks (David Bau) 09:40 AM - 10:00 AM: Keynote 1: Lee Sharkey: "Mech Interp: Where should we go from here?" 10:00 AM - 10:10 A... New England Mechanistic Interpretability Workshop

We are presenting our NNsight / NDIF demos at NEMI now!

Tune in:
youtube.com/live/q8Su4C...

NEMI 2025

The NEMI conference is live!

Watch our livestream here: youtube.com/live/q8Su4C...

About:The New England Mechanistic Interpretability (NEMI) workshop aims to bring together academic and industry researchers from the New England and surround... New England Mechanistic Interpretability Workshop

This Friday NEMI 2025 is at Northeastern in Boston, 8 talks, 24 roundtables, 90 posters; 200+ attendees. Thanks to
goodfire.ai/ for sponsoring! nemiconf.github.io/summer25/

If you can't make it in person, the livestream will be here:
www.youtube.com/live/4BJBis...

NDIF Team We're a research computing project cracking open the mysteries inside large-scale AI systems. The NSF National Deep Inference Fabric consists of a unique combination of hardware and software that provides a remotely-accessible computing resource for scientists and students to perform detailed and reproducible experiments on large pretrained AI models, such as open large language models. We aim to make AI interpretability research more accessible through this channel by publishing lectures and educational content covering real interpretability research.

Announcing a deep net interpretability talk series!

Every week you will find new talks on recent research in the science of neural networks. The first few are posted: jackmerullo.bsky.social, Roy Rinberg, and me.

At the @ndif-team.bsky.social Youtube Channel: www.youtube.com/@NDIFTeam

NDIF Team We're a research computing project cracking open the mysteries inside large-scale AI systems. The NSF National Deep Inference Fabric consists of a unique combination of hardware and software that provides a remotely-accessible computing resource for scientists and students to perform detailed and reproducible experiments on large pretrained AI models, such as open large language models. We aim to make AI interpretability research more accessible through this channel by publishing lectures and educational content covering real interpretability research.

Subscribe to our channel for future updates:
www.youtube.com/@NDIFTeam

We will use this channel to post lectures on AI interpretability research, educational information, NDIF and NNsight updates, and more. If you're interested in collaborating on a video or would like to suggest a topic, please reach out!

David Bau is an Assistant Professor of Computer Science at Northeastern University's Khoury College. His lab studies the structure and interpretation of deep... ROME: Locating and Editing Factual Associations in GPT with David Bau

Our YouTube channel is live! Our first video features @davidbau.bsky.social‬ presenting the ROME project:
www.youtube.com/watch?v=eKd...

Want to try it for yourself? Check out our new mini-paper tutorial in NNsight to see how intervening on concept induction heads can reveal language-invariant concepts and cause a model to paraphrase text!

🔗 nnsight.net/notebooks/m...

Using causal mediation analysis on words that span multiple tokens, @sfeucht.bsky.social et al. found concept induction heads that are separate from token induction heads.

🔗 dualroute.baulab.info/

Induction heads are attention heads that help complete patterns by copying tokens (transformer-circuits.pub/2021/framew...), but can they also copy over concepts?

Great to present what’s coming next for NDIF at the @actinterp.bsky.social workshop at #ICML2025!

If you missed us, let’s chat after the conference. Reach out here: forms.gle/LtTyYnkaxDyg...

NNsight Applied Example Suggestion We welcome suggestions for tutorials using the NNsight library to explore model internals! If you would like to share an idea for an NNsight mini paper tutorial, please fill out the form below. Note: if you wish to share an existing walkthrough notebook on the NNsight website, please go fill out this form.

We welcome suggestions and contributions!
📨 To highlight your existing notebook using NNsight on our website, submit it here: forms.gle/T8aLr1YLbBA....
💡 If you'd like us to create a tutorial for a paper or concept, suggest it here: forms.gle/KfwrtbW2j9u...

We’re collaborating with researchers in the field to provide detailed, educational, and replicable notebook tutorials of recent papers. Check out nnsight.net/applied_tut... for a current list of mini paper tutorials. We plan to release a new tutorial every week.

We’re excited to announce a new series of applied "mini paper" tutorials! The goal of this series is to help researchers get hands-on experience with findings, methods, and results from recent papers in interpretability using NNsight and NDIF.

Google Colab

Excited to share our first paper replication tutorial, walking you through the main figures from "Do Language Models Use Their Depth Efficiently?" by @robertcsordas.bsky.social

🔎 Demo on Colab: colab.research.google.com/github/ndif-...

📖 Read the full manuscript: arxiv.org/abs/2505.13898