正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Our infra lets us steer trillion-parameter frontier models in real time: - live, mid-CoT edits to internal activations - directly altering how the model reasons (not just outputs) - stackable edits - no added latency We can make models more Gen Z, more concise, etc.

Goodfire

26,877 subscribers

29,647 次观看 • 9 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

7 条评论

Goodfire 的头像

Goodfire9 个月前

Thanks for having us @gopalkraman @spc!

Keate Pulse 的头像

Keate Pulse8 个月前

Exciting developments at @GoodfireAI! Real-time steering of AI models opens up new possibilities, especially in improving responsiveness and control. These innovations could redefine user interaction with AI systems! 🤖

Himanshu Kumar 的头像

Himanshu Kumar9 个月前

The system allows for live modification of internal activations with zero added latency.

Min Chon Chi 的头像

Min Chon Chi9 个月前

Directly altering how the model reasons seems very powerful.

justin curl 的头像

justin curl9 个月前

So cool!! how do these steering features interact with jailbreak resilience? I.e., if you steer it towards a feature that's inconsistent with safeguards, does it make it easier to jailbreak?

Veeraraju Elluru 的头像

Veeraraju Elluru9 个月前

This is so good!

BuBxPrivacy 🌊 RIVER|.edge🦭 的头像

BuBxPrivacy 🌊 RIVER|.edge🦭9 个月前

这是真的在改脑子啊

相关视频

🚀 New update for the Grafilab AI ecosystem! We’ve just updated new frontier models in our Playground and Unified API, including the highly anticipated models by DeepSeek and Qwen families! 🧠⚡ But that’s not all… you can now run Side-by-Side Model Comparisons directly in the Grafilab Playground! 🥊 Compare outputs from Gemini, DeepSeek, Qwen, and more in real-time, on the exact same prompt, with zero code. Find the perfect model for your workload instantly. Try it out now:

🚀 New update for the Grafilab AI ecosystem! We’ve just updated new frontier models in our Playground and Unified API, including the highly anticipated models by DeepSeek and Qwen families! 🧠⚡ But that’s not all… you can now run Side-by-Side Model Comparisons directly in the Grafilab Playground! 🥊 Compare outputs from Gemini, DeepSeek, Qwen, and more in real-time, on the exact same prompt, with zero code. Find the perfect model for your workload instantly. Try it out now:

Grafilab - Unleash AI

805,542 次观看 • 2 个月前

Pocket TTS goes multilingual! Now you can use our 100M-parameter models to generate speech in six languages, fast enough to run real-time without a GPU. We also improved the quality of the English model while keeping the same size. And all of this is open-source.

Pocket TTS goes multilingual! Now you can use our 100M-parameter models to generate speech in six languages, fast enough to run real-time without a GPU. We also improved the quality of the English model while keeping the same size. And all of this is open-source.

kyutai

29,983 次观看 • 4 个月前

🚀 Our "technical" marketer might not be looped in, but today is our biggest launch day yet. We're introducing two new products to serve the inference lifecycle: Model APIs and Training. Model APIs are frontier models running on the Baseten Inference Stack, purpose-built for production. Baseten Training (Beta) provides infra and tooling without limitations for AI models destined for production. Huge shoutout to the many partners and customers we've worked with as we built these two new products—more details below.

🚀 Our "technical" marketer might not be looped in, but today is our biggest launch day yet. We're introducing two new products to serve the inference lifecycle: Model APIs and Training. Model APIs are frontier models running on the Baseten Inference Stack, purpose-built for production. Baseten Training (Beta) provides infra and tooling without limitations for AI models destined for production. Huge shoutout to the many partners and customers we've worked with as we built these two new products—more details below.

Baseten

35,350 次观看 • 1 年前

learned a lot from this conversation with Simon Mo and Matt Bornstein. biggest takeaways for me: -there are a lot of reasons why we should like open-weight models. a lot of these arguments stop at handwavy things like "what if the labs stop releasing frontier models to the public" or "it's lower cost." but simon's position as lead maintainer of vLLM and CEO of Inferact give him authority to talk about some of the other, more interesting and concrete reasons to pay attention to open-weight models, namely that they allow end-users to calibrate latency / other performance metrics with way more customizability than what any of the frontier closed-source labs offer (and without the fear that your job might be met with a refusal at some random point where you're deep in a 2 hour job) -re: the above point...for this reason, a lot of US companies (inferact included!) choose to use open-weight models over their closed-source alternatives. this also isn't limited to internal workloads / research - on a recent a16z podcast the team at Decagon spoke about how something like 90% of their customer service ai agents run on open-weight models that they've fine-tuned. -we should really appreciate how many companies/teams came out researchers fascinated by the wave of very small open-weight models that were being distilled from e.g. gpt-3.5 and earlier models in 2022/2023 (prior to the release of chatGPT!). these small models motivated the development of pagedattention, which then led to vlmm/inferact (at other layers of the stack with similar origin stories, you can look at teams like openrouter or ollama). in other words, we have open-weight models to thank for a bunch of the orchestration infra we now rely on. i think yet another, indirect, way we can point to open-source/weight infra pushing the frontier forward. anyway, a lot more in this convo, it was a lot of fun!

learned a lot from this conversation with Simon Mo and Matt Bornstein. biggest takeaways for me: -there are a lot of reasons why we should like open-weight models. a lot of these arguments stop at handwavy things like "what if the labs stop releasing frontier models to the public" or "it's lower cost." but simon's position as lead maintainer of vLLM and CEO of Inferact give him authority to talk about some of the other, more interesting and concrete reasons to pay attention to open-weight models, namely that they allow end-users to calibrate latency / other performance metrics with way more customizability than what any of the frontier closed-source labs offer (and without the fear that your job might be met with a refusal at some random point where you're deep in a 2 hour job) -re: the above point...for this reason, a lot of US companies (inferact included!) choose to use open-weight models over their closed-source alternatives. this also isn't limited to internal workloads / research - on a recent a16z podcast the team at Decagon spoke about how something like 90% of their customer service ai agents run on open-weight models that they've fine-tuned. -we should really appreciate how many companies/teams came out researchers fascinated by the wave of very small open-weight models that were being distilled from e.g. gpt-3.5 and earlier models in 2022/2023 (prior to the release of chatGPT!). these small models motivated the development of pagedattention, which then led to vlmm/inferact (at other layers of the stack with similar origin stories, you can look at teams like openrouter or ollama). in other words, we have open-weight models to thank for a bunch of the orchestration infra we now rely on. i think yet another, indirect, way we can point to open-source/weight infra pushing the frontier forward. anyway, a lot more in this convo, it was a lot of fun!

Elena

12,922 次观看 • 1 个月前

Sam on Kimi, distillation, and open source: “I have always assumed that there are going to be great cheap models in the world, and we better be the greatest and the cheapest. You get a better deal today, at least at a particular latency, using OpenAI's models than Kimi. We distill our own models. That’s how we make smaller, cheaper models. That’s a very good thing to do. There will clearly be an important place for open source models in the world and people that will want their own weights. We have so much usage of our models that we do not need to be a gigantically high-margin business to be able to afford model training. I would rather people not distill from us, for sure. But this is not in my top ten list of worries.”

Sam on Kimi, distillation, and open source: “I have always assumed that there are going to be great cheap models in the world, and we better be the greatest and the cheapest. You get a better deal today, at least at a particular latency, using OpenAI's models than Kimi. We distill our own models. That’s how we make smaller, cheaper models. That’s a very good thing to do. There will clearly be an important place for open source models in the world and people that will want their own weights. We have so much usage of our models that we do not need to be a gigantically high-margin business to be able to afford model training. I would rather people not distill from us, for sure. But this is not in my top ten list of worries.”

Patrick OShaughnessy

269,663 次观看 • 1 个月前

Adobe + Topaz Labs 🤝🙌 Just announced—you can now access select Topaz models directly in Photoshop inside your workflow. We’ve been working hard to make this happen, and now we get to share it. 🙌It's all live in Photoshop right now. See the thread for more details.

Adobe + Topaz Labs 🤝🙌 Just announced—you can now access select Topaz models directly in Photoshop inside your workflow. We’ve been working hard to make this happen, and now we get to share it. 🙌It's all live in Photoshop right now. See the thread for more details.

Topaz Labs

53,263 次观看 • 10 个月前

At Cohere we deploy our models directly to our customers, instead of them sending data to us. It makes our job harder, but their business more secure: “When you’re using a consumer app, they are using the data you’re putting in there, training on it, all the time” - Nick Frosst

At Cohere we deploy our models directly to our customers, instead of them sending data to us. It makes our job harder, but their business more secure: “When you’re using a consumer app, they are using the data you’re putting in there, training on it, all the time” - Nick Frosst

Cohere

10,460 次观看 • 2 个月前

Today we're announcing Base Labs, a dedicated research organization focused on advancing open-source AI. We believe in a healthy, open frontier model ecosystem. To enable this, we are working on: - Blue-sky research on continual learning, the science of RL, and how models learn across their full lifecycle, with every experiment and recipe shared openly. - The BaseHub Data Foundry: the highest-quality open RL environments, training data, and real-world benchmarks, built for anyone to train and benchmark on. - Post-post training: taking open-source models and making them better, safer, and more aligned through continual post-training, built on our research, and deployed with our frontier safety stack so organizations can use open-source models with confidence. - Making models cheaper and more performant through our model performance research. This is a mission-driven research effort, not a commercial product. We believe the health of the open-source AI ecosystem matters and that the best way to advance it is to do science in the open. We’re hiring engineers, researchers, and research fellows to advance this mission.

Today we're announcing Base Labs, a dedicated research organization focused on advancing open-source AI. We believe in a healthy, open frontier model ecosystem. To enable this, we are working on: - Blue-sky research on continual learning, the science of RL, and how models learn across their full lifecycle, with every experiment and recipe shared openly. - The BaseHub Data Foundry: the highest-quality open RL environments, training data, and real-world benchmarks, built for anyone to train and benchmark on. - Post-post training: taking open-source models and making them better, safer, and more aligned through continual post-training, built on our research, and deployed with our frontier safety stack so organizations can use open-source models with confidence. - Making models cheaper and more performant through our model performance research. This is a mission-driven research effort, not a commercial product. We believe the health of the open-source AI ecosystem matters and that the best way to advance it is to do science in the open. We’re hiring engineers, researchers, and research fellows to advance this mission.

Base Labs

232,099 次观看 • 13 天前

The more AI can do, the more we need to ask what it should and shouldn’t do. OpenAI researcher Jason Wolfe joins host Andrew Mayne to explore the Model Spec, the public framework that defines how models are intended to behave. They break down how it works in practice, from the chain of command that resolves conflicting instructions to the way it evolves over time through real-world use, feedback, and new model capabilities.

The more AI can do, the more we need to ask what it should and shouldn’t do. OpenAI researcher Jason Wolfe joins host Andrew Mayne to explore the Model Spec, the public framework that defines how models are intended to behave. They break down how it works in practice, from the chain of command that resolves conflicting instructions to the way it evolves over time through real-world use, feedback, and new model capabilities.

OpenAI

225,959 次观看 • 5 个月前

Speech-native models like Moshi sound great and answer fast, but aren’t as smart as text LLMs. In our new paper, MoshiRAG, we show how Moshi can ask for advice from a text LLM or a knowledge base. The tricky part is how to do this in real time without adding latency. 🧵

Speech-native models like Moshi sound great and answer fast, but aren’t as smart as text LLMs. In our new paper, MoshiRAG, we show how Moshi can ask for advice from a text LLM or a knowledge base. The tricky part is how to do this in real time without adding latency. 🧵

kyutai

53,284 次观看 • 4 个月前

LLM-grounded Diffusion: Enhancing Prompt Understanding of Text-to-Image Diffusion Models with Large Language Models paper page: github: Recent advancements in text-to-image generation with diffusion models have yielded remarkable results synthesizing highly realistic and diverse images. However, these models still encounter difficulties when generating images from prompts that demand spatial or common sense reasoning. We propose to equip diffusion models with enhanced reasoning capabilities by using off-the-shelf pretrained large language models (LLMs) in a novel two-stage generation process. First, we adapt an LLM to be a text-guided layout generator through in-context learning. When provided with an image prompt, an LLM outputs a scene layout in the form of bounding boxes along with corresponding individual descriptions. Second, we steer a diffusion model with a novel controller to generate images conditioned on the layout. Both stages utilize frozen pretrained models without any LLM or diffusion model parameter optimization. We validate the superiority of our design by demonstrating its ability to outperform the base diffusion model in accurately generating images according to prompts that necessitate both language and spatial reasoning. Additionally, our method naturally allows dialog-based scene specification and is able to handle prompts in a language that is not well-supported by the underlying diffusion model.

LLM-grounded Diffusion: Enhancing Prompt Understanding of Text-to-Image Diffusion Models with Large Language Models paper page: github: Recent advancements in text-to-image generation with diffusion models have yielded remarkable results synthesizing highly realistic and diverse images. However, these models still encounter difficulties when generating images from prompts that demand spatial or common sense reasoning. We propose to equip diffusion models with enhanced reasoning capabilities by using off-the-shelf pretrained large language models (LLMs) in a novel two-stage generation process. First, we adapt an LLM to be a text-guided layout generator through in-context learning. When provided with an image prompt, an LLM outputs a scene layout in the form of bounding boxes along with corresponding individual descriptions. Second, we steer a diffusion model with a novel controller to generate images conditioned on the layout. Both stages utilize frozen pretrained models without any LLM or diffusion model parameter optimization. We validate the superiority of our design by demonstrating its ability to outperform the base diffusion model in accurately generating images according to prompts that necessitate both language and spatial reasoning. Additionally, our method naturally allows dialog-based scene specification and is able to handle prompts in a language that is not well-supported by the underlying diffusion model.

AK

83,681 次观看 • 3 年前

Decart just dropped Lucy 2.5 and it completely changes what live video can do. This is their Live AI model that edits video as it streams, no post production needed. > add, replace, or remove anything mid stream > restyle entire scenes on the fly > generate realtime VFX with a single prompt > better prompt adherence and more realism than before > already live across streaming, e-commerce, and advertising We are officially in the era where AI does not just generate video, it directs it live.

Decart just dropped Lucy 2.5 and it completely changes what live video can do. This is their Live AI model that edits video as it streams, no post production needed. > add, replace, or remove anything mid stream > restyle entire scenes on the fly > generate realtime VFX with a single prompt > better prompt adherence and more realism than before > already live across streaming, e-commerce, and advertising We are officially in the era where AI does not just generate video, it directs it live.

Abdul Șhakoor

207,955 次观看 • 2 个月前

Why the largest enterprises will not use Chinese models and how regulation will enforce that "Given the regulatory environment in the US, it's probably more likely that we see a great American open-source competitor arise. We're going to have at least one massive, multi-hundred-billion-dollar, if not trillion-dollar, American company focused on American-first open source. The largest enterprises will want an American open source alternative." Anastasios Nikolas Angelopoulos What are large US enterprises more scared of; frontier models competing with them or Chinese models and data security? Jason Warner Eiso Kant Josh Wolfe Horace He Martin Ziqiao Ma John Schulman

Why the largest enterprises will not use Chinese models and how regulation will enforce that "Given the regulatory environment in the US, it's probably more likely that we see a great American open-source competitor arise. We're going to have at least one massive, multi-hundred-billion-dollar, if not trillion-dollar, American company focused on American-first open source. The largest enterprises will want an American open source alternative." Anastasios Nikolas Angelopoulos What are large US enterprises more scared of; frontier models competing with them or Chinese models and data security? Jason Warner Eiso Kant Josh Wolfe Horace He Martin Ziqiao Ma John Schulman

Harry Stebbings

19,299 次观看 • 1 个月前

Kimi K3 full fine-tuning is live on AC2. Our memory optimizations reduced GPUs required per training replica by ~40%. At nearly 3T parameters, Kimi forced us to rethink how we manage memory, communication, rollouts, and checkpoints. The result is a much more efficient path to training frontier-scale open models.

Kimi K3 full fine-tuning is live on AC2. Our memory optimizations reduced GPUs required per training replica by ~40%. At nearly 3T parameters, Kimi forced us to rethink how we manage memory, communication, rollouts, and checkpoints. The result is a much more efficient path to training frontier-scale open models.

Applied Compute

234,012 次观看 • 18 天前

Introducing OpenVenice — an open-source, customizable frontend for Venice. Your API key stays in your browser, talks directly to Venice. No server, no backend. - Chat, image gen, audio, music, video, embeddings — everything venice offers in one clean UI - Visual workflows — chain models together (vision → LLM → image gen -> video gen). Venice is multimodal so you can build some really powerful workflows - Full control over every model parameter - Runs locally, deploys anywhere — zero infrastructure I'm a $DIEM holder and heavy Venice user. The models are solid and website is great but I kept wanting more freedom with the UX — more control, custom tools, being able to give it access to my local computer, multimodal chaining -- something I could shape to how I actually work. Didn't exist, so I built it. Feel free to fork and make it yours. Open source, MIT licensed.

Introducing OpenVenice — an open-source, customizable frontend for Venice. Your API key stays in your browser, talks directly to Venice. No server, no backend. - Chat, image gen, audio, music, video, embeddings — everything venice offers in one clean UI - Visual workflows — chain models together (vision → LLM → image gen -> video gen). Venice is multimodal so you can build some really powerful workflows - Full control over every model parameter - Runs locally, deploys anywhere — zero infrastructure I'm a $DIEM holder and heavy Venice user. The models are solid and website is great but I kept wanting more freedom with the UX — more control, custom tools, being able to give it access to my local computer, multimodal chaining -- something I could shape to how I actually work. Didn't exist, so I built it. Feel free to fork and make it yours. Open source, MIT licensed.

nikshep

40,500 次观看 • 6 个月前

From Palantir and Two Sigma to turning mechanistic interpretability into a production-grade platform, Mark Bissell (Member of Technical Staff) and Myra Deng (Head of Product) are building Goodfire on the belief that “understanding what’s inside the model” is the path to safer, more controllable AI and Goodfire just raised a $150M Series B at a $1.25B valuation to scale that bet. We sat down with Mark and Myra to unpack what actionable interpretability looks like beyond demos: why SAE features can lose to raw-activation probes on real detection tasks, how they deployed token-level PII monitoring with Rakuten across English + Japanese without training on real customer data, and what it takes to do real-time steering on a trillion-parameter model (including the infamous Gen-Z feature). We also dig into hallucination detection, surgical post-training edits for bias/unlearning, the emerging equivalence between activation steering and in-context learning, and their push to bring interpretability into training so model development isn’t just “data in, weights out.” Goodfire mark bissell Myra Deng

From Palantir and Two Sigma to turning mechanistic interpretability into a production-grade platform, Mark Bissell (Member of Technical Staff) and Myra Deng (Head of Product) are building Goodfire on the belief that “understanding what’s inside the model” is the path to safer, more controllable AI and Goodfire just raised a $150M Series B at a $1.25B valuation to scale that bet. We sat down with Mark and Myra to unpack what actionable interpretability looks like beyond demos: why SAE features can lose to raw-activation probes on real detection tasks, how they deployed token-level PII monitoring with Rakuten across English + Japanese without training on real customer data, and what it takes to do real-time steering on a trillion-parameter model (including the infamous Gen-Z feature). We also dig into hallucination detection, surgical post-training edits for bias/unlearning, the emerging equivalence between activation steering and in-context learning, and their push to bring interpretability into training so model development isn’t just “data in, weights out.” Goodfire mark bissell Myra Deng

Latent.Space

19,389 次观看 • 7 个月前

How smoking every day year after year affects your lungs visualized in 3D using GPT 6 Astra 🫁 Nah, that’s not 10 different models there or any shady dark magic. It’s just ONE model that I built using Tripo. Astra worked directly on that glb model using Three.js, added morph targets, and used TSL shaders to show how smoking affects the lungs in real time. The whole model is also 84% compressed using Meshopt and WebP compression. I feel like experiences like this can make it much easier for students to understand what smoking actually does to your lungs over time and maybe even help a few people quit. This is how AI can be put to good use, imo. Live:

How smoking every day year after year affects your lungs visualized in 3D using GPT 6 Astra 🫁 Nah, that’s not 10 different models there or any shady dark magic. It’s just ONE model that I built using Tripo. Astra worked directly on that glb model using Three.js, added morph targets, and used TSL shaders to show how smoking affects the lungs in real time. The whole model is also 84% compressed using Meshopt and WebP compression. I feel like experiences like this can make it much easier for students to understand what smoking actually does to your lungs over time and maybe even help a few people quit. This is how AI can be put to good use, imo. Live:

The Bugged Dev

71,990 次观看 • 8 天前

Monitor and control your AI spend on every provider on Our early users save 40% on average. Every week, the price-intelligence-latency frontier shifts, and we expect this trend to continue. Tradeoffs between latency, reasoning, cost, service tier, open source and closed source models are shifting constantly. Router sends every request to the model that's actually best for the task and helps you control what tokens you buy. We benchmark it against real work: ~40% lower cost for the same outputs. Today we're opening it to everyone. Two lines of code or just change your base URL. No Ramp account needed. Free through 2026, first $26 on us. Get an API key today at

Monitor and control your AI spend on every provider on Our early users save 40% on average. Every week, the price-intelligence-latency frontier shifts, and we expect this trend to continue. Tradeoffs between latency, reasoning, cost, service tier, open source and closed source models are shifting constantly. Router sends every request to the model that's actually best for the task and helps you control what tokens you buy. We benchmark it against real work: ~40% lower cost for the same outputs. Today we're opening it to everyone. Two lines of code or just change your base URL. No Ramp account needed. Free through 2026, first $26 on us. Get an API key today at

Veeral Patel

1,421,256 次观看 • 27 天前

We are opening up the beta of our new AI writing tool to Exclusive Models. If you're exclusive to JFF, try it out and let us know your thoughts. At our Anniversary Conference, lots of models said they needed help making their posts more creative. This is a tool to help models take what they write and help them make better posts. Does this replace a human? No! In fact, if you are an erotic writer, apply to be in our Business Directory so models can hire you to help them out!

Sensitive content

We are opening up the beta of our new AI writing tool to Exclusive Models. If you're exclusive to JFF, try it out and let us know your thoughts. At our Anniversary Conference, lots of models said they needed help making their posts more creative. This is a tool to help models take what they write and help them make better posts. Does this replace a human? No! In fact, if you are an erotic writer, apply to be in our Business Directory so models can hire you to help them out!

JustForFans

15,642 次观看 • 3 年前

Too many blockchain infra providers run in the cloud or on rented servers That's a problem is built different 3 reasons why you want to work with us: 1) Best in class, enterprise grade bare metal, all fully owned 2) SOC 2 Type II certified--not just the data centers, but our own internal team 3) Optimized performance on our infra for your staked assets Curious to learn more? Check out the video and short overview thread below!

Too many blockchain infra providers run in the cloud or on rented servers That's a problem is built different 3 reasons why you want to work with us: 1) Best in class, enterprise grade bare metal, all fully owned 2) SOC 2 Type II certified--not just the data centers, but our own internal team 3) Optimized performance on our infra for your staked assets Curious to learn more? Check out the video and short overview thread below!

Ryan (“Phunky”) Haczynski

39,192 次观看 • 1 年前