The First Mechanistic Interpretability Frontier Lab — Myra Deng & Mark Bissell of Goodfire AI
Latent Space: The AI Engineer PodcastEpisode Description
Summary
This episode features Mark Bissell and Myra Deng from Goodfire AI, an AI research lab focused on mechanistic interpretability, discussing their recent $150M Series B funding round and their work in applying interpretability to real-world production scenarios. They delve into challenges with Sparse Autoencoders (SAEs), demonstrate real-time model steering on a trillion-parameter model, and highlight applications in scientific discovery and healthcare for understanding and controlling AI behavior.
Chapters
Topics
People
Key Concepts (13)
References (30)
Transcript (126 segments)
So welcome to the Lanespace five. We're back in the studio with our special Mech Interp co host, Vibu. Welcome.
Mochi. Mochi's special co host. And Mochi taco.
Mechanistic interpretability doggo. We have with us Mark and Myra from Goodfire. Welcome.
Thanks for having us on. Maybe we can sort of introduce Goodfire and then and then introduce you guys. How do you introduce Goodfire today?
Yeah. It's a great question. So Goodfire, we like to say is an AI research lab that focuses on using interpretability to understand, learn from and design AI models.
And we really believe that interpretability will unlock the new generation next frontier of safe and powerful AI models. That's our description right now. And I'm excited to dive more into the work we're doing to make that happen.
Yeah. And there's there's always like the official description. Is there an, like, an unofficial one that sort of resonates more with a different audience?
being an AI research lab that's focused on interpretability, there's obviously a lot of people have a lot that they think about when they think of interpretability. And I think we have a pretty broad definition of what that means and the types of places that can be applied. And in particular, applying it in production scenarios, in high stakes industries, and really taking it sort of from the research world into the real world, which, you know, it's an it's a new field, so that hasn't been done all that much, we're excited about actually seeing that sort of put into put into practice.
Yeah. I would say it's it wasn't too long ago that Anthropic was, like, still putting out, like, toy models, a superposition, and that kind of stuff. And I wouldn't have pegged it to be this far along.
When you and I talked at NeurIPS, you were talking a little bit about your production use cases and your customers. And then not to bury the lead, today we're also announcing the fundraise, your series b, a 150,000,000 million dollars at a 1.25.
Congrats, you're a unicorn. Thank you. Yeah.
No. Things move fast. We were talking to you in December and Yeah.
Already some big updates since then. Let's dive, I guess, into a bit of your backgrounds as well. Mark, you you were at Palantir working on health stuff, which is really interesting because the Goodfire has some interesting, like, health use cases.
I don't know how related they are in practice.
Yeah.
I don't know. It was helpful context to know what it's like just to work with, health systems and generally in that domain. Yeah.
And, Mara, you were at two Sigma, which actually I was also at, two Sigma Oh, really? Back in the day. Wow.
Nice. Did we overlap at all? No.
and now you head of product. What are your sort of respective roles just to introduce people to, like, what all gets done in Goodfire? Yeah.
Prior to Goodfire, was at Palantir for about three years as a as a forward deployed engineer, now a now a hot term. I wasn't always that way, and as a technical lead on the health care team. And at Goodfire, I'm a a member of the technical staff.
And honestly, that I think is about as specific as I could describe myself because I've worked on a range of things and, you know, it's a fun time to be at a team that's still reasonably small. Think when I joined one of the first 10 employees, now we're above 40, but still it looks like there's always a mix of research and engineering and product and all of the above that needs to get done. I think everyone across the team is pretty switch hitter in the roles they do.
I think you've seen some of the stuff that I worked on related to image models, which was a research demo. More recently, I've been working on our scientific discovery team with some of our life sciences partners, but then also building out our core platform for flexing some of the kind of MLE and and developer skills as well. Very generalist.
And you you also had like a very like a founding engineer type role. Yeah. Yeah.
So I I also started as I still am a member of technical staff, did a wide range of things from the very beginning including, like, finding our office space and all of these nitty gritty deals. We both we both visited when you had that open house thing. It was really nice.
Thank you. Thank you. Yeah.
Plugged to come visit our office.
It it looked like it was it was like 200 people. Like, it has room for 200 people, but you guys are like 10.
a while, it was very empty.
But, yeah, like like Mark, I I spend a lot of my time as as head of product. I think product is a bit of a weird role these days. But a lot of it is thinking about how do we take our frontier research and really apply it to the most important real world problems and how does that then translate into a platform that's repeatable or a product and working across, you know, the engineering and research teams to make that happen.
And also communicating to the world, like, what is interpretability?
What is it used for? What is it good for? Why is it so important?
All of these things are part of my day to day as well. I love like what is things because that's a very crisp like starting point for people like coming to a field. Maybe I'll do a fun thing.
Vibhu, why you wanna try tackling what is interpretability and then they can correct us. Okay. Great.
So I think, like, one, just to kick off, it's a very interesting role to be head of product. Right? Because you guys, at least as a lab, you're more of an applied interp lab.
Right? Which is pretty different than just normal interp, like a lot of background research, but you guys actually ship an API to try these things. You have Ember, you have products around it, which not many do.
Okay. What is interp? So basically, you're trying to have an understanding of what's going on in model, like in the model, in the internal.
So different approaches to do that. You can do probing, SAEs, transcoders, all this stuff. But basically you have an you have a hypothesis.
You have something that you wanna learn about what's happening in a model internals, and then you're trying to solve that. From there, you can do stuff like you can, you know, you can do activation mapping, you can try to do steering.
what's happening on the model internals. How'd I do? That was really good.
I know it's great. I think it's also a, it's kind of a minefield of a, if you ask 50 people who quote unquote work in Interp, like, what is interpretability? You'll probably get 50 different answers, and to some extent also, like, where- where Goodfire sits in the space, I think that we're an AI research company above all else, and interpretability is a is a set of methods that we think are really useful and worth kind of specializing in, in order to accomplish the goals we want to accomplish.
But I think we also sort of see some of the goals as even more broader as as almost like the science of deep learning and just taking a not black box approach to kind of any part of the, like, AI development life cycle, whether that means using Interp for, like, data curation while you're training your model or for understanding what happened during post training or for the, you know, understanding activations and sort of internal representations, what is in there semantically. And then a lot of sort of exciting updates that were, you know, are sort of also part of the the fundraise around bringing interpretability to training, which I don't think has been done all that much before. A lot of this stuff is sort of post hoc poking at models as opposed to actually using this to intentionally design them.
Is this post training or pre training? Or is is that not a useful post training, but there's no reason the techniques wouldn't also work in pre training. Yeah.
basically, I I'm thinking like rollouts or like, you know, having different variations of a model that you can tweak with the Exactly. Steering. Yeah.
And I think in a lot of the news that you've seen in in on like Twitter or whatever, you've seen a lot of unintended side effects come out of post training processes, you know, overly sycophantic models or models that exhibit strange reward hacking behavior. I think these are like extreme examples. There's also, you know, very mundane, more mundane, like enterprise use cases where, you know, they try to customize or post train a model to do something and it learns some noise or it doesn't appropriately learn the target task.
And a big question that we've always had is like, how do you use your understanding of what the model knows and what it's doing to actually guide the learning process more effectively?
Yeah. I mean, you know, just to anchor this for people, one of the biggest controversies of last year was four o Glaze Gate. Never heard of Glaze Gate.
I didn't know that was what it was called. No. The other one did call it that on the blog post.
And I was like, why did OpenAI call it, like, officially use that term? And I'm like, that's funny. But like, yeah, I I I guess it's it's the pitch that if they had worked a good fire, they wouldn't have avoided it.
Like, you know, I'd say.
I think so. Yeah. I think that's certainly one of the use cases.
I think another reason why post training is a place where this makes a lot of sense is a lot of what we're talking about is surgical edits. You know, you want to be able to have expert feedback very surgically change how your model is doing, whether that is, you know, removing a certain behavior that it has. So, you know, one of the things that we've been looking at or is is another, like, common area where you would want to make a somewhat surgical edit is some of the models that have, say, political bias.
bias in them and, you know Is there a CCP vector?
Well, there's there's certainly internal, yeah, parts of the representation space where you can sort of see where that lives. Yeah.
And you wanna kinda, you know, extract that piece out. Well, I always say, you know, whenever you find a vector, a fun exercise is just, like, make it very negative Yeah. To see what's what the opposite of CCP is.
super America, bald eagles flying everywhere. But yeah. So in general, like, lots of post training tasks where you'd wanna be able to to do that.
Whether it's unlearning a certain behavior or, you know, some of the other kind of cases where this comes up is do are you familiar with, like, the the grokking behavior?
I mean, I know the machine learning term of grokking.
Yeah. Sort of this, like, double the scent idea of of having a model that is able to learn a generalizing a generalizing solution as opposed to even if memorization of some task would suffice, you want it to learn the more general way of doing a thing. And so, you know, another way that you can think about having surgical access to a model's internals would be learn from this data, but learn in the right way, if there are many possible, you know, ways to to do that.
Can Mekinterb solve the double descent problem? Depends, I guess, on how you Okay. So I I I view double descent as a problem.
Uh-huh. Because then you're like, well, if the loss curves level out, then you're done, but maybe you're not done. Right.
Right. But, like, if you actually can interpret what is generalizing or what is is still changing even though the loss is not changing, then maybe you you can actually not view it as a double the set problem and actually you're just kind of translating the space in which you view loss and you you like, and then you have a smooth curve. Yeah.
I I think that's certainly like the domain of of problems that we're that we're looking to get. Yeah. To me, like, double descent is like the biggest thing to like ML research where, like, if you believe in scaling, then you don't you need to know where to scale and but if you believe in double the standard, don't you don't believe in anything where, like, anything levels off.
Yeah. I mean, also tangentially, there's like okay. When you talk about the China vector, right, there's the subliminal learning work.
It was from the Anthropic Fellows program where basically you can have hidden biases in a model. And as you distill down or, you know, as you train on distilled data, those biases always show up even if, like, you explicitly try to not train on them. So, you know, it's just like another use case of, okay.
If we can interpret what's happening in post training, you know, can we clear some of this? Can we even determine what's there? Because, yeah, it's just like some worrying research that's out there that shows, you know, we really don't know what's going on.
That is yeah. I think that's the biggest sentiment that we're sort of hoping to tackle. Nobody knows what's going on.
Right? Like, subliminal learning is just an insane concept when you think about it. Right?
Train a model on not even the logits, literal the literally the output text of a bunch of random numbers, and now your model loves OWLs.
they defy they defy intuition, and and there are mathematical explanations
that you can get into, but mean, they're early days. Objectively, there are a sequence of numbers that are more OL like than others. There there should be.
to according to certain models. Right? It's interesting.
starting So usually, yes.
like, probably it will transfer across different models so. I think of it more as a statistical artifact of models initialized from the same seed sort of. There's something that is, like, path dependent from that seed that might cause certain overlaps in the latent space and then sort of doing this distillation, yeah, like, pushes it towards having certain other tendencies.
Got it. I think there's more research. A bunch of these open ended questions.
Right? Like, you can't train in new stuff during the RL phase. Right?
RL only reorganizes weights, and you can only do stuff that's somewhat there in your base model. You're not learning new stuff. You're just reordering chains and stuff.
But okay. My broader question is, when you guys work at an Interp lab, how do you decide what to work on and what's kind of the thought process? Right?
Because we can ramble for hours. Okay. I wanna know this.
I wanna know that. But, like, how do you concretely like, know, what's the workflow? Okay.
There's, like, approaches towards solving a problem. Right? I can try prompting.
I can look at chain of thought. I can train probes, SAEs. But how do you determine, you know, like, okay, is this going anywhere?
Like, do we have set stuff? Just, you know, if can talk about that. It's a really good question.
at the very beginning of the company, thought about like, let's go and try to learn what isn't working in machine learning today. Whether that's talking to customers or talking to researchers at other labs, trying to understand both where the frontier is going and where things are are really not falling apart today. And then developing a perspective on how we can push the frontier using interpretability methods.
And so, you know, even our chief scientist, Tom, spends a lot of time talking to customers and trying to understand what real world problems are and then taking that back and trying to apply the current state of the art to those problems and then seeing where they fall basically. And then using that those failures or those shortcomings to understand what hills to climb when it comes to interpretability research. So, like on the fundamental side, for instance, when we have done some work applying SAEs and probes, we've encountered, you know, some shortcomings in SAEs that we found a little bit surprising.
And so have gone back to the drawing board and done work on better foundational interpreter models. And a lot of our team's research is focused on what is the next evolution beyond SAEs, for instance. And then when it comes to like control and design of models, you know, we tried steering with our first API and realized that it still fell short of black box techniques like prompting or fine tuning.
So And went back to the drawing board and were like, how do we make that not the case? And how do we improve it beyond that? And one of our researchers, Ekdeep, who just joined is actually Ekdeep and Atticus are like steering experts and and have spent a lot of time trying to figure out like, what is the research that enables us to actually do this in a much more powerful, robust way.
is, like, look at real world problems, try to translate that into a research agenda, and then, like, hill climb on on both of those at the same time. Yeah. Mark has the steering CLI demo queued up, which we're gonna go into a sec.
But I always wanna double click on when you drop hints like we found some problems of SAEs.
Okay. What are they? You know?
And let let's let's then we can then we can go into the demo. Yeah. I mean, I am curious if you have more thoughts here as well because you've done it in in the health care domain.
But I think like, for instance, when we do things like trying to detect behaviors within models that are harmful or like behaviors that user might not want to have in their model. So hallucinations, for instance, harmful intent, PII, all of these things. We first tried using SAE probes for a lot of these tasks.
So taking the feature activation space from SAEs and then training class classifiers on top of that, and then seeing how well we can detect the properties that we might want to detect in model behavior. And we've seen in many cases that probes just trained on raw activations seem to perform better than SAE probes, which is a bit surprising if you think that SAEs are actually also capturing the concepts that you would want to capture cleanly and more surgically. And so that is an interesting observation.
I don't think that is like I'm not down on SAEs at all. I think there are many, many things they're useful for. But we have definitely run into cases where I think the concept space described by SAEs is not as clean and accurate as as we would expect it to be for actual, like, real world downstream performance metrics.
Fair enough. Yeah. Yeah.
It's the blessing and the curse of unsupervised methods where you get to peek into the AI's mind, but sometimes you wish that you saw other things when you when you went inside there.
I think weren't an SAE based approach actually did prove to be the most It did work well in in the case that we published with Rakuten. And I think a lot of the reasons it worked well is because we had a noisier dataset. And so actually, the blessing of unsupervised learning is that we actually got to get more meaningful generalizable signal from SAEs when the data was noisy.
But in other cases where we've had like good datasets, it hasn't been the case.
usage or sort of production usage? Yeah. So they are using us to essentially guardrail and inference time monitor their language model usage and their agent usage to detect things like PII so that they don't route private user information to downstream model providers.
And so that's, you know, going through all of their user queries every day. And that's something that we deployed with them a few months ago. And now we are actually exploring very early partnerships, not just with Rakuten, but with other people around how we can help with potentially training and customization use cases as well.
Yeah.
in Japan. Yes. Yeah.
And I think that use case actually highlights a lot of like what it looks like to deploy things in practice that you don't always think about when you're doing sort of research tasks. So when you think about some of the stuff that came up there that's more complex than your idealized version of a problem, they were encountering things like synthetic to real transfer of methods. So they couldn't train probes, classifiers, things like that on actual customer data of PII.
So what they had to do is use synthetic datasets and then hope that that transfer is out of domain to real datasets, and so we can evaluate performance on the real datasets but not not train on customer PII. So that right off the bat is like a big challenge. You have multilingual requirements, so this needed to work for both English and Japanese text.
Japanese text has all sorts of quirks, including tokenization behaviors that caused lots of bugs that caused us to be pulling our hair out. And then also a lot of tasks you'll see, you might make simplifying assumptions if you're sort of treating it as, like, the easiest version of the problem to just sort of get, like, general results where maybe you say you're classifying a sentence to say, this contain PII? But the need that Rakuten had was token level classification so that you could precisely scrub out the PII.
So as we learned more about the problem, you're sort of speaking about what that looks like in practice. Yeah. A lot of assumptions end up breaking, and that was just one instance where you a problem that seems simple right off the bat ends up being more complex as you keep diving into it.
Excellent. One of the things that's also interesting with Interp is a lot of these methods are very efficient. Right?
compared to a separate, like, guardrail, LM as a judge, a separate model. One, have to host it. Two, there's, like, a whole latency.
So if you use, a big model, you have a second call.
it's also deployed for efficiency. Right? So thinking of someone like Rakuten doing it in production live, you know, that's just another thing people should consider.
Yeah. And something like a probe is super lightweight. Yeah.
It's no extra latency, really. Excellent. You have the steering nimbles lined up, so we were just kinda see what you got.
thing. No. This is a pretty hacky demo from from a presentation that someone else on the team recently gave.
So this will give a sense for for steering in action. Honestly, think the biggest thing that this highlights is that as we've been growing as a company and taking on kind of more and more ambitious versions of interpretability related problems, a lot of that comes to scaling up in various different forms. And so here you're going to see steering on a 1,000,000,000,000 parameter model.
This is Kimi K2. And so it's sort of fun that in addition to the research challenges, there are engineering challenges that we're now tackling. Because for any of this to be sort of useful in production, you need to be thinking about what it looks like when you're using these methods on frontier models, as opposed to sort of like toy kind of model organisms.
So, yeah, this was thrown together hastily, pretty fragile behind the scenes, but I think it's quite a fun demo. So screen sharing is Is on? Is on.
So I've got two terminal sessions pulled up here. On the left is a forked version that we have of the Kimi, CLI that we've got running to point at our custom hosted Kimi model. And then on the right is a setup that will allow us to steer on certain concepts.
So I should be able to chat with Kimi over here. Tell it hello. Is this running locally?
So the CLI is running locally, but the Kimi server is running back in the office. Well, hopefully, it should be. That's too much to run on that Mac.
Yeah. I think it's, it takes a full, like, h 100 node. I think it's, like, you can run it on eight GPUs, h 100.
So so, yeah, Kimi's running. We can ask it a prompt. It's got a forked version of our, of the SGLAN code base that we've been working on.
So I'm gonna tell it, hey, this SGLAN code base is slow. I think there's a bug. Can you try to figure it out?
It's a big code base, so it'll it'll spend some time doing this. And then on the right here, I'm gonna initialize in real time some steering. Let's see here.
Continue searching for any bugs. Feature ID forty two zero five. Players twenty thirty forty.
So let me this is basically a feature that we found that inside Kimi seems to cause it to speak in Gen Z slang. And so on the left, it's still sort of thinking normally. It might take, I don't know, fifteen seconds for this kick in, but then we're gonna start hopefully seeing it.
Dude, this code base is massive for real. So we're gonna start seeing Kimmy transition as the steering kicks in from normal Kimmy to Gen Z Kimmy. And both in its chain of thought and its actual outputs.
And interestingly, you can see, you know, it's still able to call tools and stuff. It's purely sort of its its demeanor. And there are other features that we found for interesting things like concision.
So that's more of a practical one. You can make it more concise. The types of programs programming languages it uses.
pretty good output. Scheduler code is actually wild. This is Sebastian.
Yoda's code is actually insane, bro.
Be cringe, Yoda. What's the process of training an SAE on this? Or, you know, how do you label features?
I know you guys put out a pretty cool blog post about finding this, like, autonomous interp, something about how agents for interp is different than, like, coding agents. I don't know while this is spewing up. How how do we find feature forty three two zero five?
Yeah. So in this case, we our platform that we've been building out for a long time now supports all the sort of classic out of the box interp techniques that you might wanna have, like, SAE training, probing, things of that kind. I'd say the techniques for, like, vanilla SAEs are pretty well established now where you take your model that you're interpreting, run a whole bunch of data through it, gather activations, and then, yeah, pretty straightforward pipeline to train an SAE.
There are a lot of different varieties. There's top K SAEs, batch top K SAEs, normal ReLU SAEs. And then once you have your sparse features, to your point, assigning labels to them to actually understand that this is a Gen Z feature, that's actually where a lot of the kind of magic happens.
And the most basic standard technique is look at all of your data input dataset examples that cause this feature to fire most highly, and then you can usually pick out a pattern. So for this feature, if I've run a diverse enough dataset through my model, feature 43,205 probably tends to fire on all the, tokens that sound like Gen Z slang. And, so, you know, you could have a human go through all 43,000 concepts and look at the pattern, but to automate that, you just kind of hand those examples off to a Frontier LLM and ask it to identify that pattern.
And I've gotta ask the basic question. You know?
pass it through, see what feature activates for hallucinations? Can I just, you know, turn hallucination down? Oh, wow.
You you really I solved it.
using interpretability techniques. And this is interesting because hallucinations is something that's very hard to detect. And it's like a kind of a hairy problem and something that black box methods really struggle with.
Whereas like Gen Z, you could always train a classify, a simple classifier to detect that. Hallucinations is harder. But we've seen that models internally have some awareness of like uncertainty or some sort of like user pleasing behavior that leads to hallucinatory behavior.
working on mitigating the hallucinatory behavior in the model itself as well. Yeah. I would say most people are still at the level of like, oh, I would just turn temperature to zero and that turns off hallucination.
And I'm like, well, that's a fundamental misunderstanding of Right. How this works. Yeah.
Although I so I part of what I like about that question is you there are SAE based approaches that might, like, help you get at that. But oftentimes, the beauty of SAEs and, like we said, the curse is that they're unsupervised. So when you have a behavior that you deliberately would like to remove and that's more of, like, a supervised task, often it is better to use something like probes and and specifically target the thing that you're interested in reducing as opposed to sort of like hoping that when you fragment the latent space, one of the vectors that pops out will be the thing you're interested in.
And as much as we're training an autoencoder to be sparse, we're not, like, for sure certain that, you know, we will get something that just correlates to hallucination. Right? You'll you'll probably split that up into 20 other things, and who knows what they'll be?
Of course. Right. Yeah.
So there's known sort of problems with, like, feature splitting and feature absorption. And then there's the off target effects. Right?
Ideally, you would wanna be very precise where if you reduce the hallucination feature, suddenly, maybe your model can't write creatively anymore and maybe you don't like that, but you wanna still stop it from hallucinating facts and figures.
Good. So Vivo has a paper to recommend there that we'll put in the show notes. But, yeah, I I mean, I guess, just because your demo is done, any any other things that you wanna highlight or any other interesting features you want to show?
I don't think so. Yeah. Like I said, this is a pretty small snippet.
I think the main sort of point here that I think is exciting is that there's not a whole lot of interp being applied to models quite at at this scale. You know, Anthropic certainly has some some research and, yeah, other other teams as well, but it's it's nice to see these techniques, you know, being put into practice.
would have sounded Yeah. The fact that it's real time, like, you you started the thing and then you edited the steering vector. I think it's it's an interesting one.
TBD, what the actual, like, production use case would be on that, like, real time editing? It's like, that's the fun part of the demo. Right?
You can kind of see how this could be served behind an API. Right? Like Yes.
You're you only have so many knobs and you can just tweak it a bit more. I don't know how it plays in. Like, people haven't done that much with like, how does this work with or without prompting.
Right? How does this work with fine tuning? Like, there's a whole hype of continual learning.
Right? So there's just so much to see. Like, is this another parameter?
Like, is it, like, parameter? We just kinda leave it as a default. We don't use it.
So don't know. Maybe someone here wants to put out a guide on, like, how to use this with prompting, when to do what.
Well, I have a paper recommendation that I think you would love from Ekdeep on our team who is an amazing researcher, just can't say enough amazing things about Ekdeep, but he actually has a paper that as well as some, you know, others from the from the team and elsewhere that go into the essentially equivalents of activation steering and in context learning and how those are from a he he thinks of everything in a cognitive neuroscience Bayesian framework, but basically how you can precisely show how prompting in context learning and steering exhibit similar behaviors and even, like, get quantitative about the, like, magnitude of steering you would need to do to induce a certain amount of behavior similar to certain prompting, even for things like jailbreaks and stuff. It's a really cool paper. Are you saying steering is less powerful than prompting?
More like you can almost write a formula that tells you how to Like a one to one. Convert between the two of them. And so, like be formally equivalent actually in the in the limit.
Right. So, like, one case study of this is for jailbreaks. There I don't know.
Have you seen the stuff where you can do, like, many shot jailbreaking? You, like Yeah. Flood the context with examples of their behavior?
We and the topic put out that paper, a lot of people were like, yeah. We've been doing this, guys. Like, we're we're in papers.
Yeah.
What's in this in context learning and activation steering equivalence paper is you can, like, predict the number of examples that you will need to to put in there in order to jailbreak the model That's cool. By doing steering experiments and using this sort of, like, equivalence mapping. That's cool.
That's really cool. It's very neat. Yeah.
you know, it updates the KV cache kind of. And and like and and and then every next token inference is still, you know, the the the the sheer sum of everything all the way it's plus all the context up to date and you could, I guess, theoretically steer that with you could probably replace that with your steering. The only problem is steering typically is on one layer, maybe three layers like like you did.
So it's like not exactly equivalent. Right. Right.
get precise about, yeah, like, how you sort of define steering and, like, what how how you're modeling the setup. But, yeah, I've got the paper pulled up here, belief dynamics reveal the dual nature. Yeah.
The title is belief dynamics reveal the dual nature of in context learning and activation steering.
Dana Wargraft on the, who are, doing fellowships at Goodfire. Egg Deep's the the final author there. I think actually to your question of like, what is the production use case of steering?
I think maybe if you just think like one level beyond steering as it is today, like imagine if you could adapt your model to be an expert legal reasoner, like in almost real time, like very efficiently using human feedback or using like your semantic understanding of what the model knows and where it knows that behavior. I think that while it's not clear what the product is at the end of the day, it's clearly very valuable. Thinking about like what's the next interface for for model customization and adaptation is interesting problem for us.
Like we have heard a lot of people actually interested in fine tuning an RL for open weight models in production. And so people are using things like Tinker or kind of like open source libraries to do that. But it's still very difficult to get models fine tuned in RL for exactly what you want them to do unless you're an expert at at model training.
And so that's like something we're looking into.
Yeah. I never thought so Tinker from Thinking Machines famously uses rank one LoRa. Is that basically the same as steering?
Like, oh, you know, what's the comparison there? Well, so in that case, you are still applying updates to the parameters. Right?
Not you're not touching a base model. You're you're touching an adapter. It's kind of yeah.
Right. But I I guess it's still is, like, more in parameter space than I guess it's maybe, like, are you modifying the pipes or are you modifying the water flowing through the pipes Okay. To get what you're after.
Yeah. Just maybe one way. Yeah.
I like that analogy. That that's my mental That's my one. At least.
But it gets at this idea of model design and intentional design, which is something that we're that we're very focused on. Yeah. And just the fact that, like, I hope that we look back at how we're currently training models and post training models and just think what a primitive way of doing that right now.
Like, there's no intentionality really in It's just data. Right? The only thing can control is what data we feed in.
So so Dan from Goodfire likes to use this analogy of, you know, he has a couple of young kids and he talks about, like, what if I could only teach my kids how to be good people by giving them cookies or, like, you know, giving them a slap on the wrist if they do something wrong? Like, not telling them why it was wrong or, like, what they should have done differently or something like that. Just Yeah.
Figure it out. The t g. Right.
Exactly. So that's RL. Yeah.
Right. And then p you know, it's sample inefficient. There's, you know, what do they say?
It's like slurping feedback. It's like Super strong. Through slurping supervision through a Right.
And so we'd like to get to the point where you can have experts giving feedback to their models that are internalized. Equality. Yeah.
And and, you know, steering is an inference time way of sort of getting that idea.
you're moving to a a world where it is much more intentional design in perpetuity for these models. K. This is one of the questions we asked Emmanuel from Anthropic on the podcast a few months ago.
Basically, the question was you're at a research lab that does model training, foundation models, and you're on an interp team. How does it tie back? Right?
Like, does this do ideas come from the pretraining team? Do they go back? You know?
So for those interested, you can you can watch that. There wasn't too much of a connect there, but it's still something. You know?
It's something they wanna push for down the line.
for all of the above. Like, there are certainly post hoc use cases where it doesn't need to touch that. I think the other thing a lot of people forget is this stuff isn't too computationally expensive.
Right? Like, I would say if you're interested in getting into research, MEKINTERP is one of the most approachable fields. Right?
A lot of this train in SAE, train a probe, this stuff like, the budget for this, one, there's already a lot done. There's a lot of open source work. You guys have done some too.
You know, you think There's, like, notebooks from the Gemini team, from Neil Nanda Right. Or, like, this this is how you do it. Just step through the notebook.
Even if you're, like, not even technical with any of this, you can still make, like, progress there. You can look at different activations.
if you do wanna get into training, you know, training this stuff, correct me if I'm wrong, is, like, in the thousands of dollars. Not even, like it's not that high scale. And then same with, like, you know, applying it, doing it for post training, RR.
All this stuff is fairly cheap in scale of, okay. I wanna get into, like, model training. I don't have compute for, like, you know, pre training stuff.
So it's it's a very nice field to get into. And also, there's a lot of, like, open questions. Right?
There's so many questions we have. Some of them have to go with, okay. I want a product.
I wanna solve this. Like, there's also just a lot of open ended stuff that people could work on that's interesting. Right?
I don't know if you guys have any calls for, like, what's open questions? What's open work that you either open collaboration with or like you just like to see solved? Or just, you know, for people listening that wanna get into MechInterpret because people always talk about it.
What are what are things they should check out? Start, of course, you know, join you guys as well. I'm sure you're hiring.
Lee, Sharkey, it's Open Problems and Interpretability, which I recommend everyone who's interested in the field read. Just like a really comprehensive overview of what are the things that experts in the field think are the most important problems to be solved. I also think to your point, it's been really, really inspiring to see, I think a lot of young people getting interested in interpretability.
Actually, not just young people, also like scientists who have been, you know, experts in physics for many years and in biology or things like this transitioning into inter because the barrier to entry is is, you know, in in some ways low and and there's a lot of information out there and ways to get started.
which was not the case a few years ago. So it just goes to to show how, I guess, like exciting the field is, how fast it's moving, how quick it is to get started, and things like that. And also just a very welcoming community.
You know, there's an open source MacInterp Slack channel. There are people who always posting questions and just folks in the space are always responsive if you ask things on various forums and stuff.
paper is is a really good one. For other people who wanna get started, I think, you know, MATS is a great program. What's What's the acronym for?
Machine Learning and Alignment Theory Scholars? Yeah.
the
normally summer internship style. Yeah. But they've doing it year round now.
And actually, a lot of our full time staff have come through that program or gone through that program. And it's great for anyone who is transitioning into interpretability. There's a couple other fellows programs we do on as well as Anthropic.
And so those are great places to get started if anyone is is interested.
as as it does scale up. I should mention that Lee actually works with you guys. Right?
And in the London office and amending our first ever Mekinterp track in at AIE Europe because I I see this industry applications now emerging, and I'm pretty excited to, you know, help get push that along. Yeah.
conference. Yeah.
I'm so glad you added that. You know, it's it's a little still a little bit of a bet. Like, there's this it's not like that widespread, but, like, I I can definitely see, like, this is the time to, like, really get into it.
Like, we we wanna be early on things. For sure. And I think the field is it understands this.
Right?
of the MEK Interp workshop this year was actionable interpretability and there was a lot of discussion around bringing it to to various domains. Everyone's everyone's adding like pragmatic. Pragmatic.
Actionable. Right. Like, whatever.
It looks like, okay. Well, we weren't actionable before, I guess. I don't know.
And I mean, like, just, you know, being at Europe, you see the interp room. One, like, old school conferences, like, I think they had a very tiny room till they got lucky and they got it doubled, but there's definitely a lot of interest. A lot of niche niche research, so you see a lot of research coming out of university students.
We covered the paper last week. It's like two unknown authors, not many citations, but, you know, you can make a lot of meaningful work there. One thing I I did want to call out because I think people haven't really mentioned this yet is just interfer code.
I think it's it's like an abnormally important field, we haven't mentioned this yet. The conspiracy theory last two years ago was when the first SAE work came out of Anthropic was they would do it like, oh, we just used SAEs to turn the bad code vector down and then turn up And then what code came out. Turn up the good code.
And I think, like, isn't that the the dream? Like, you know, like but basically, I I guess, maybe why is it funny? Like, it's it's if it was realistic, it would not be funny.
It will be like, no, actually, we should do this. But it's funny because we know there's like we feel there's some limitations to what steering can do. And I think a lot of the public image of steering is like the Gen Z stuff, like like, oh, you can make it really love the Golden Gate Bridge or you can make it speak like Gen Z to like be a legal reasoner seems like a huge stretch.
Yeah. And I don't know if that will get there this way. Yeah.
I will say we are announcing something very soon that I will not speak too much about. But I think, yeah, this is like what we run into again and again is like, we we don't want to be in the world where steering is only useful for like stylistic things. Things that's definitely not not what we're aiming for.
But I think the types of interventions that you need to do to get to things like legal reasoning are much more sophisticated and require breakthroughs in learning algorithms and that's And is this an emergent property of scale as well? I think so. Yeah.
I mean, think scale definitely helps. I think scale allows you to learn a lot of information and and reduce noise across, you know, large amounts of data. But I also think, we think that there's ways to do things much more effectively even even at scale.
So like actually learning exactly what you want from the data and not learning things that you do that you don't want exhibited in the data. So we're not like anti scale, but we are also realizing that scale is not going to get us to the type of AI development that we want to be at in in the future as these models get more powerful and get deployed in all these sorts of like mission critical contexts. Current life cycle of training and deploying and evaluations is is to us like deeply broken and has opportunities to to improve.
So more to come on that very, very soon.
if you can manipulate them in the precise best way, you can get the ideal combination of them that you desire. And steering is maybe the most coarse grained sort of peek at what that looks like, but I think it's evocative of what you could do if you had total surgical control over every concept of Every each parameter. Yeah.
Exactly.
code features. I've got it pulled up. Yeah.
Yeah. Just coincidentally, as you guys were talking to them. This is like this is exactly what people have thought it.
There's, like, specifically a code error feature to activate. So just they show it off. It's not it's not typo detection.
It's like it's it's typos in code. It's not typical typos. And, you know, you can you can see it clearly activates where there's something wrong in code.
And they have, like, malicious code, code error. They have a whole bunch of sub, you know, sub broken down little grain features.
Yeah. Yeah. So so the the rough intuition for me, the why I talked about post training was that, well, you just, you know, have a few different rollouts with all these things turned off and on and whatever, and then, you know, you can that's that's synthetic data you can kinda post train on.
Yeah.
the real hard work. I mean, guys have the right idea exactly.
Yeah, we replicated a lot of these features in our LAMA models as well. Remember there was like And I think a lot of this stuff is open, right? Like, yeah, you guys opened yours.
DeepMind has opened a lot of essays on Gemma. Even Anthropic has opened a lot of this. There's there's a lot of resources that, you know, we can probably share Yeah.
Of people that wanna get involved. Yeah. And special shout out to, like, Neuronpedia as well.
Mhmm. Yes.
amazing piece of work to visualize those things. Yeah. Exactly.
I guess I wanted to pivot a little bit on onto the health care side because I think that's a big use case for you guys. We haven't really talked about it yet. This is a bit of a crossover for me because we are set we are we do have a separate science pod that we're starting up for AI for science.
Wow. Okay. Just because like it's such a huge investment category.
And also I'm like less qualified to do it. We actually have bio PhD's to cover that, which is great. But I think that you just kind of reco recap your work maybe on the evil two stuff, but then and then building forward.
Yeah. For sure.
I think another kind of interesting just lens on interpretability in general is a lot of the techniques that we're describing are ways to solve the AI human interface problem. And it's sort of like bidirectional communication is the goal there. So what we've been talking about with intentional design of models and, you know, steering, but also more advanced techniques, is having humans impart our desires and control into models and over models.
And the reverse is also very interesting, especially as you get to superhuman models, whether that's narrow superintelligence, like these scientific models that work on genomics data, medical imaging, things like that. But down the line, you know, super intelligence of other forms as well, what knowledge can the AIs teach us as sort of that the the other direction in that? And so some of our life science work to date has been getting at exactly that question, which is, well, some of it does look like debugging these various life sciences models, understanding if they're actually performing well on tasks or if they're picking up on spurious correlations.
For instance, genomics models, you would like to know whether they are sort of focusing on the biologically relevant things that you care about or if it's using some simpler correlate like the ancestry of the person that it's looking at. But then also in the instances where they are superhuman and maybe they are understanding elements of the human genome that we don't have names for or, specific, you know, yeah, discoveries that they've made that that we don't know about, that's that's a big goal. And so We're already seeing that.
We are partnered with organizations like Mayo Clinic, a leading research health system in The United States, Arc Institute, as well as a startup called PrimaMenta, which focuses on neurodegenerative disease. In our partnership with them, we've used foundation models they've been training and applied our interpretability techniques to find novel biomarkers for Alzheimer's disease.
that's like a flavor of some of the things that we're working on. Yeah. I think that's really fantastic.
We obviously, we did the Chad Zuckerberg pod last year as well, and, like, there's gonna be a plethora of these models coming out because there's so much potential and research. And it's like very interesting how it's basically the same as language models, but just with a different underlying dataset. But, like, it's the same exact techniques.
Like, there's no change, basically. Yeah.
Well, and even in, like, other domains. Right? Like, I I you know, robotics, I know, like, a lot of the companies just use Gemma as, like, the the, like, backbone, and then they, like, make it into a VLA that, like, takes these actions.
It's transformers all the way down. Yeah, other two things.
have Medgema now, right? Like this week, even there was Medgema 1.5 and they're training it on this stuff, like three d scans, medical domain knowledge and all that stuff too.
So there's a push from both sides. But I think the thing that, you know, one of the things about MechInterp is like, you're a little bit more cautious in some domains. Right?
So healthcare mainly being one, like guardrails understanding, you know, we're we're more risk adverse to something going wrong there. So even just from a basic understanding, like if we're trusting these systems to make claims, we wanna know why and what's going on. Yeah.
to actually using foundation models for real patient usage or things like that. Like say you're using a model for rare disease prediction, you probably want some explanation as to why your model predicted a certain outcome and an interpretable explanation at that. So that's definitely a use case.
But I also think like being able to extract scientific information that no human knows to accelerate drug discovery and disease treatment and things like that actually is a really, really big unlock for science, like scientific discovery. And you've seen a lot of startups like say that they're going to accelerate scientific discovery. And I feel like we actually are doing that through our interp techniques.
And kind of like almost by accident, like I think we got reached out to very, very early on from these healthcare institutions and none of us had healthcare How did they even hear of you? A podcast.
Okay.
Podcast. Okay.
for can call us up. Podcasts are the most important thing. Everyone should listen to podcasts.
Everyone should come on to podcasts. They were like, you know, we have these really smart models that we've trained and we wanna know what they're doing. And we were like really early that time, like three months old and it was a few of us and we were like, oh my god.
Looks like we've never used these models. Let's figure it out. But it's also like great proof that interp techniques scale pretty well across across domains.
We didn't really have to learn too much about. Interp is a machine learning technique. Machine learning skills everywhere.
Right? Like, and then there's obviously Exactly.
insight. Yeah. Probably to finance too, which would be fun for Mhmm.
Our history. Mhmm. I don't know if you have anything to say there.
Yeah. Well, just across the side, like we've also done work on material science.
Yeah. It it really runs the gamut. Yeah.
Awesome. And, you know, for those that should reach out, like, obviously experts in this, but like, is there a call out for people that you're looking to partner with, design partners, people to use your stuff outside of just, you know, the general developer that wants to plug and play steering stuff, like on the research side more so, like, are there ideal design partners, customers, stuff like that? That's just Yeah.
I can talk about maybe non life sciences and then I'm curious to hear from you on the life sciences side.
many domains, language, anyone who's customizing language models or trying to push the frontier of code or reasoning models is really interesting to us. And then also interested in the frontier of models that work in like pixel space as we call it. So if you're doing world models, video models, even robotics, where there's not a very clean natural language interface to interact with.
I think we think that Interp can really help and are looking for a few partners in that space.
Just because you mentioned the keyword world models. Is that a big part of your thinking? Do you have a definition that I can use because everyone's asking me about it?
About world models? There's quite a few definitions, let's say. I don't feel equipped to be an expert on world model definitions.
But the reason we're interested in them is because they give you like, you know, with language models, when you get features, you still have to do auto interpret and things like that to actually get an understanding of of what this concept is. But in image and video and world, it's like extremely easy to to grok what the concept is because you can see it and you can visualize it. And this makes the feedback cycle extremely fast for us.
And also for things like, I don't know, if you take think about probes in language model context and then take it to world models. Like what if you wanted to detect harmful actors in world model scenes? Like you can't actually like go and label all of that data feasibly, but maybe you could synthetically generate, you know, I don't know, world like harmful actor data using SAE feature activations or whatever.
And then actually train a probe that was able to detect that much more scalably. So I just think like video and image and world has always been something we've explored and are continuing to explore.
this could really, like, change the world of The steering demo? Yeah. No.
That image demo. The the diffusion one. Yeah.
The fusion. Exactly. Yeah.
We should probably show that. And and you demoed it at World's Fair. So Yeah.
We can share the link that. Nice. Yeah.
People can play with it. Right? Yes.
Yeah. Still have. I think for for me, one way in which I I think about world models is just like this, like, having this consistent model of the world where everything that you generate operates within the rules of that world.
And imagine it would be a bigger deal for science or, like, math or anything that where, like, you have verifiable rules. Whereas, I guess, in natural language, maybe there's less rules. And so it's not not that important.
Yeah.
debugging of the model's internal representations or its internal world model to the extent you can make that legible and explicit and have control over that, I think it makes it all the more important. Because in language, it's sort of a fuzzy enough domain that Yeah. If its world model isn't fully like ours, it can still sort of, like, pass the Turing test, so to speak.
But I know there have been papers that have looked at, like, even if you train certain astrophysics models, it does not learn f equals m a. Like, the same way that you can, you know, have a model do well for modular arithmetic, but it doesn't really, like, learn what how we think of modular arithmetic. It learns some crazy heuristic that is, like, essentially functionally equivalent, but it's probably not the sort of Grokt solution that you would hope for.
It's how an alien would do it. Right.
Right. You know? Right.
Exactly. But but no. No.
I think there's probably, I think, a function of our learning being bad rather than the well, that approach probably not being real because it's it's how we humans learn. Right. Yeah.
Right. Well, it's just it's the problem of induction. Right?
All of ML is based on induction, and and it's impossible to say I have a physics model.
when there is a character wearing a blue shirt and green shoes, and, like, you you can't disprove that that's the case unless you test every particular situation your model might be in. Yeah. So we know that the laws of physics apply no matter where you are, what scenario it is.
But from a model's perspective, maybe something that's out of distribution, it just never needed to learn that the same laws of physics apply there. Yeah.
and I was very inspired by this short story called Understand, which apparently is like pretty old. You you must be familiar with it. To me, it was like it's it's this fictional story.
It's like the inverse of Florence for Algernon where you had someone like get really smart, but then also try to outsmart the tester. And the story just read like the chain of thought of a of a super intelligence. Yeah.
Right? Where they're like, oh, I realize I'm being tested. Therefore and then, okay, what's the consequence of being tested?
Oh, they're testing me and if I score well, they will use me for things that I don't wanna do, therefore I will score badly and I and I like but not too badly that they will raise alarm. So this is like so model sandbagging is a thing that people have have explored.
Ted Chiang's work just in general seems to be something that inspires you. I just wanted to prompt you to talk about it. I think so Ted Chiang has two is a is a sci fi author who writes amazing short stories and His other claim to fame is Stories of Our Lives, which became, the movie Arrival.
Exactly. Yeah. So so two books of short stories that I'm aware of.
He also actually also has a great, just online blog post. I think he's the one who coined the term of LLMs as, like, a blurry JPEG of the Internet. I should fact check that, but he it's a good it's a good post.
But I think almost every one of his short stories has some lesson to bear on thinking about AI and thinking about AI research. So, you know, you you've been talking about alien intelligence, right, and this AI human communication translation problem. That's, you know, exactly sort of what's going on in Arrival in Story of Story of Your Life.
And just the fact that other beings will think and operate and communicate in ways that are not just challenging for us to understand, but just fundamentally different in ways that we might not even be able to expect. And then the one that's just super relevant for interpretability is the other short book of short stories he has is called exhalation Yes. And that is literally about a robot doing interpretability on its own mind.
Oh, okay. So I just think that that, you know, you don't you don't even have to squint to make the analogies there.
Well, I I actually think exhalation is a discussion about entropy and order. But, yes, there's this there's a scene in exhalation where basically the the and so everyone is a robot. So the the the guy realizes he can set up a mirror to work on the back of his own head and then starts doing operations like that and by looking at the mirror and doing doing this.
Yeah. And I think Ted Chiang has written about, like, the inspiration for that story. It was, like, half inspired by some of the thing he had been doing on Entropy.
There's apparently some other short story that is similar where a character goes to the doctor and opens up his chest and there's like a like a ticker tape going along. He is like he basically realizes he's like a Turing machine. And I don't know.
think especially as it comes to using agents for Interp, that story always sticks in my mind. I find the brain surgery or like surgery analogies a little bit a little bit morbid, but it is very apt. And when we talk to a lot of computational neuroscientists, they moved to Interp because they were like, look, we have unfettered access to this artificial intelligent mind.
It's so much You have access to everything. You can run as many ablations experiments as you want. It's an amazing bed for science.
And human brains, obviously we can't just go and do whatever we want to them. And I think it is really just like a moment in time where we have intelligent systems that can really like do things better than humans in many ways. And it's time, I think, for us to to do the science on it.
I'll ask a a brief like safety question.
Mac and Terrible was kind of born out of the alignment and safety conversation. Safety is on your website. It's not like something that you you like deprioritize, but like there's like a sort of very militant safety arm that like wants to blow up data centers and like stop AI and and then there's this like sort of middle ground and like is is this like a conversation in your part of the world?
Do you go up to Berkeley and Light Haven and like talk to those guys or are they like, you know, there's like a brief like civil war going on? I don't know.
good amount of us have spent some time in Berkeley. And then there are researchers there that we really admire and respect. I think for us, it's like we have a very grounded view of alignment and safety in that we want to make sure that we can build models that do what we want them to do and that we have scalable oversight into what these models are doing.
And and we think that that is the key to a lot of these like technical alignment challenges. And I think that is our opinion. That's our research direction.
We of course are going to do safety related research to make sure that our techniques also work on, you know, things like reward hacking and other like more concrete safety issues that we've seen in the wild. But we want to be kind of like grounded in solving the technical challenges we see to having humans be humans play a big role in in the deployment of of these super intelligent agents of the future.
Yeah. I've I've found the community to actually be remarkably cohesive, whether it's talking about academia or the interpretability work being done at the Frontier Labs or some of the independent programs like maths and stuff, I think we're all shooting for the same goal. I don't know that there's anyone who doesn't want our understanding of models to increase.
I I think everyone, regardless of where they're coming from or the use cases that they're thinking, whether it's alignment as the premier thing they're focused on or someone who's coming in purely from the angle of scientific discovery. I think we would all hope that models can be more reliably and robustly controlled and understood. It seems like a pretty unambiguous goal.
I'll maybe phrase it in terms of like, there's maybe like a u curve of of this where like, if you're extremely doomer, you don't want any research whatsoever. If you're like, mildly doomer, you're like, okay, there's there's like high agency doomer is like, well, the default path is we're we're all dead, but like, we can do something about it.
you know? Yeah. Yeah.
There's also the other side. Like, there is the super alignment, like people that are like, okay, weak to strong generalization. We're gonna get there.
We're gonna have models smarter than us and use those to train even smarter models. How do we do that safely? That's, you know, there's the camp there too.
That's trying to solve it. But, there's there's a lot of doomers too.
Well, and I and I think there's a lot to be learned from taking a very, like, even regardless of the problems that you're applying this to, also just, like, the notion of, like, scalable oversight as a method of saying, let's take super intelligent or or current frontier models and help use them to understand other models is another case where I think it's just, a good lesson that everyone is aligned on of, ideally, you are setting up your research so that as super intelligence arrives, that is a tailwind that's also bolstering our ability to, like, understand the models. Because otherwise, you're fighting a losing battle if it's like the systems are getting more and more capable and our methods are sort of linearly growing at, like, human pace.
Yeah. Yeah. Vivo got did call out something like, you know, I I I do think a consistent part of the MEK INTERP field is consistently strong to weak, meaning that we we train weaker models to understand strong models, something like that.
Or maybe I got it the other way around. The other way, weak to Yeah. Yeah.
The question that Ilya and Yanlaika posed was, well, does that gonna scale? Because eventually these are gonna be stronger than us. Right?
So I I don't know if you have a perspective on that because I that is something I still haven't got over. But after seeing that There's a good paper from OpenAI, but it's somewhat old. I think it's like '23, '24.
It's literally weak to strong generalization. Yeah. But the thing is that most of OpenAI's super alignment team has They're gone.
They're gone. But like, I think the idea the idea is solid. It's it's bad.
There's no more They're still back? I I think there's some new blog posts coming out of it. I know.
Did just, you know, check the thinking machines website to see who's back. They're still back. There's more kind of thing.
You know you know what mean? Like, We Too Strong seemed like a very different direction. And when when it first came out, I was like, oh my god, this is this is what we have to do.
Mhmm. And like, it may be completely different than everything, all all the techniques that we have today.
My understanding of that is it's that's more like weak to strong when you when you trust the weak model and you're uncertain whether you can trust the strong model that's that's being developed. I'm sort of speaking out of my depth on some of these topics, but I think right now we're in a regime where even the strong models we trust is reasonably aligned, and so they can be good co scientists on a lot of the problems that we've been we've been tackling, which is a nice a nice state to be in. Yeah.
Any last thoughts? Call section?
I don't think so. As we mentioned, actively hiring MLEs, research scientists. You can check out the careers page at Goodfire.
Where are you guys based?
San Francisco. We're in Levi's Plaza, like by Court Tower. It's where our office is.
So come hang out. We're also looking for design partners across people working in reasoning models, world models, robotics. And then also, of course, people who are working on building super intelligent science models or looking at drug discovery or disease treatment, we would love to partner as well.
maybe you have a use case where LLMs are almost good enough, but you need one magical knob to tune so that it is good enough that you guys make the knob. Yeah. Yeah.
in in other domains as well. The the some of those are the, especially opaque ones because you can't you can't chat with them. So, like, what do you what do you do if you can't chat with them?
Oh, well, like, thinking about, like, a genomics model or a material science model. So, like, a yeah.
model. Yeah. They predict.
Yeah. Got it. Got it.
I was just gonna I thought the diffusion work you guys did early was pretty, you know, pretty fun.
diffusion or images. Right? Like, I see genomics Oh, it's gonna be huge.
Like, look at this video models. They're so expensive to produce. And, like and I mean, basically, the mid journey s ref is kind of a feature.
Right? The what? Mid journey s ref.
Oh, yeah. The the the string of numbers that you Right. Right.
Right. Yeah. The style reference, I guess.
Yeah. No. I I mean, I think we're starting to see more of it.
And I'll say, like, the the research preview of our diffusion model, kind of like a creative use case in the steering demo you saw, I I think of those much more as as as demos than, a lot of the sort of core platform features that that we're working with partners are unfortunately sort of under NDA and less demo able, but I will, you know, hope that you're gonna see Inter pervading a lot of what gets done even if it is behind the scenes like that. So some of the, yeah, some of the public facing demos might not always be representative of, like, the it's it's just the tip of the iceberg, I guess, is one way to put it. Okay.
Excellent. Thanks for coming on. Thanks for having us.
This is a great time.
Shared via Hopper