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。

LLM Visualization This is actually pretty amazing! It helps to visualize the core components of LLMs like nano-gpt and GPT-3.

elvis

299,907 subscribers

39,431 次观看 • 1 年前 •via X (Twitter)

教育科学技术新闻政治

Anya Rossi• Live Now

Private livecam show

9 条评论

AI Advocate Arif 的头像

AI Advocate Arif1 年前

I'd love to see more visualizations like this make complex LLM concepts more accessible

Mehrdad Yazdani 的头像

Mehrdad Yazdani1 年前

Does anyone actually find visualizations useful? I don’t mean to hate, but besides educational purposes is there any other use? Code/equations are so much easier for me to understand.

elvis 的头像

elvis1 年前

Fair comment. I think, and as you said, they are pretty useful for educational purposes. I don't know that this particular one can lead to any new research insights as is. I really liked the other Transformer Explainer visualization as it does show more things like the transformation of the data, probs, and so on.

通往AGI之路的头像

通往AGI之路1 年前

Mind blown 🤯

elbouz 的头像

elbouz1 年前

It's insane how easier this makes it to understand complex architecture. Also great for efficient refreshers.

GPT.Biz 的头像

GPT.Biz1 年前

This looks like a great resource to understand LLMs better, definitely worth checking out!

Clancy 的头像

Clancy1 年前

That is awesome

AIxBlock 的头像

AIxBlock1 年前

Nice breakdown and visualization! Tks so much for sharing 👏

Rajesh David 的头像

Rajesh David1 年前

Truly amazing stuff. You know what could be cooler ? Going step-by-step and assembling them part-by-part with some details about why that is needed ? Does it already do that ?

相关视频

This is the inside of a GPT/LLM as it reads your text and "thinks" about what to say next

This is the inside of a GPT/LLM as it reads your text and "thinks" about what to say next

Brandon

43,353 次观看 • 1 年前

Ever wonder what an LLM actually looks like under the hood? This visualization is mesmerizing. Check it out yourself:

Ever wonder what an LLM actually looks like under the hood? This visualization is mesmerizing. Check it out yourself:

Matthew Berman

166,215 次观看 • 2 年前

Sam Altman says the leap from GPT-4 to GPT-5 will be as big as that of GPT-3 to 4 and the plan is to integrate the GPT and o series of models into one model that can do everything

Sam Altman says the leap from GPT-4 to GPT-5 will be as big as that of GPT-3 to 4 and the plan is to integrate the GPT and o series of models into one model that can do everything

Tsarathustra

224,102 次观看 • 1 年前

Sam Altman says GPT-3 was the first real glimpse of passing a spiritual Turing test GPT-5 shows the first small signs of AI doing new science — useful ideas, real help on papers GPT-6 will be a leap like GPT-3 to 4, but this time for science, where it can really do it

Sam Altman says GPT-3 was the first real glimpse of passing a spiritual Turing test GPT-5 shows the first small signs of AI doing new science — useful ideas, real help on papers GPT-6 will be a leap like GPT-3 to 4, but this time for science, where it can really do it

Haider.

163,642 次观看 • 8 个月前

分享一个可深入了解大语言模型 LLM 的工作原理网站：LLM Visualization。该网站通过生动的 3D 可视化交互方式，清晰展示 LLM 推理过程所有步骤，让我们更加直观学习 LLM 背后的工作原理。链接：目前已提供了 GPT-2(small)、nano-gpt、GPT-2(XL) 和 GPT-3 模型的工作原理可视化。

分享一个可深入了解大语言模型 LLM 的工作原理网站：LLM Visualization。该网站通过生动的 3D 可视化交互方式，清晰展示 LLM 推理过程所有步骤，让我们更加直观学习 LLM 背后的工作原理。链接：目前已提供了 GPT-2(small)、nano-gpt、GPT-2(XL) 和 GPT-3 模型的工作原理可视化。

GitHubDaily

13,304 次观看 • 1 年前

Sam Altman says the leap from GPT-4 to GPT-5 will be as big as that of GPT-3 to 4 and the plan is to integrate the GPT and o series of models into one model that can do everything - that’s the AGI

Sam Altman says the leap from GPT-4 to GPT-5 will be as big as that of GPT-3 to 4 and the plan is to integrate the GPT and o series of models into one model that can do everything - that’s the AGI

Chubby♨️

231,778 次观看 • 1 年前

OpenAI's Brad Lightcap: "o1 is almost like a portal to GPT-7, GPT-8... [it] gives you the effective compute of what a GPT-7 or GPT-8 would... it's a pure discontinuity on the scaling"

OpenAI's Brad Lightcap: "o1 is almost like a portal to GPT-7, GPT-8... [it] gives you the effective compute of what a GPT-7 or GPT-8 would... it's a pure discontinuity on the scaling"

Tsarathustra

72,679 次观看 • 1 年前

GPT-5.6 vs GPT-5.5 on my custom spaceship prompt. I gave both models the exact same custom prompt. This is also the same prompt I previously gave to Fable 5. For context, GPT-5.6 Pro worked for 87 minutes, while GPT-5.5 Extra High worked for 34 minutes and 42 seconds. As I’ve said before, based on great authority GPT-5.6 will be an incremental/soldi improvement over GPT-5.5, not a “Fable killer.” My rough expectation has been that it would trade blows with Fable 5 on some benchmarks, maybe win around half depending on the category, but not clearly surpass it overall. And again fable five will have bigger model smell, but this was expected. After testing this coding output, that view feels pretty accurate. GPT-5.6 is clearly better than GPT-5.5 in several visual areas. The lighting, shading, chairs, object details, and exterior of the spaceship looked noticeably stronger. The scene was also easier to test. I do want to give GPT-5.5 credit though. It built out the rooms much much better and the planets looked better than GPT-5.6’s. It was also interesting that both GPT-5.5 and GPT-5.6 produced better-looking planets than Fable 5 in this specific test. The downside with GPT-5.5 was stability. The game was much glitchier and harder to test compared to GPT-5.6. But when it comes to the core of the demo, which is the spaceship itself, Fable 5 still beat both models pretty comfortably. GPT-5.6 is impressive, but from this test, it looks exactly like what I expected which was a meaningful incremental improvement over GPT-5.5, at least for indie game demos, but not something that replaces Fable 5. In collaboration with Chetaslua

GPT-5.6 vs GPT-5.5 on my custom spaceship prompt. I gave both models the exact same custom prompt. This is also the same prompt I previously gave to Fable 5. For context, GPT-5.6 Pro worked for 87 minutes, while GPT-5.5 Extra High worked for 34 minutes and 42 seconds. As I’ve said before, based on great authority GPT-5.6 will be an incremental/soldi improvement over GPT-5.5, not a “Fable killer.” My rough expectation has been that it would trade blows with Fable 5 on some benchmarks, maybe win around half depending on the category, but not clearly surpass it overall. And again fable five will have bigger model smell, but this was expected. After testing this coding output, that view feels pretty accurate. GPT-5.6 is clearly better than GPT-5.5 in several visual areas. The lighting, shading, chairs, object details, and exterior of the spaceship looked noticeably stronger. The scene was also easier to test. I do want to give GPT-5.5 credit though. It built out the rooms much much better and the planets looked better than GPT-5.6’s. It was also interesting that both GPT-5.5 and GPT-5.6 produced better-looking planets than Fable 5 in this specific test. The downside with GPT-5.5 was stability. The game was much glitchier and harder to test compared to GPT-5.6. But when it comes to the core of the demo, which is the spaceship itself, Fable 5 still beat both models pretty comfortably. GPT-5.6 is impressive, but from this test, it looks exactly like what I expected which was a meaningful incremental improvement over GPT-5.5, at least for indie game demos, but not something that replaces Fable 5. In collaboration with Chetaslua

Chris

228,126 次观看 • 18 天前

Try out GPT-V in Cursor! It's pretty good for building/modifying components!

Try out GPT-V in Cursor! It's pretty good for building/modifying components!

Aman Sanger

195,264 次观看 • 2 年前

The Email Marketing Brain 🧠 This Custom GPT I just finished training and developing with 500 pages of knowledge is INSANE. It can: - Write ACTUALLY good copy - Create campaign calendars - Answer email marketing questions Like + follow + comment "GPT" and I'll send access 🤝

The Email Marketing Brain 🧠 This Custom GPT I just finished training and developing with 500 pages of knowledge is INSANE. It can: - Write ACTUALLY good copy - Create campaign calendars - Answer email marketing questions Like + follow + comment "GPT" and I'll send access 🤝

Max Sturtevant

27,034 次观看 • 1 年前

🚨 Host Your Own Local LLM On The Abacus AI SuperComputer Stop being sad about Fable and control your destiny by hosting your own LLM - host open source LLMs like Qwen and Gemma - create chat bots or always on APIs - message it via an always on agents Use SOTA models like GPT 5.5xHigh or Opus to create open-source LLM apps and games

🚨 Host Your Own Local LLM On The Abacus AI SuperComputer Stop being sad about Fable and control your destiny by hosting your own LLM - host open source LLMs like Qwen and Gemma - create chat bots or always on APIs - message it via an always on agents Use SOTA models like GPT 5.5xHigh or Opus to create open-source LLM apps and games

Bindu Reddy

18,777,375 次观看 • 23 天前

The next change mentioned in the last AMA on Discord is the addition of "Mention GPTs" and a more agent-like approach to GPTs This new feature allows you to inline tag any other custom GPT, which will then be displayed as "Talking to [Grimoire/custom GPT name]". This feature also enables using different GPTs in the same conversation, taking an agents-like approach (although it seems that this feature may not work properly yet). To mention a GPT and add it directly into your conversation, simply type "@" followed by the name of the GPT. h/t to Glenn 'devalias' Grant & his amazing ChatGPT Source Watch project

The next change mentioned in the last AMA on Discord is the addition of "Mention GPTs" and a more agent-like approach to GPTs This new feature allows you to inline tag any other custom GPT, which will then be displayed as "Talking to [Grimoire/custom GPT name]". This feature also enables using different GPTs in the same conversation, taking an agents-like approach (although it seems that this feature may not work properly yet). To mention a GPT and add it directly into your conversation, simply type "@" followed by the name of the GPT. h/t to Glenn 'devalias' Grant & his amazing ChatGPT Source Watch project

Tibor Blaho

12,776 次观看 • 2 年前

It works!! This is fully designed by GPT-5.5 and GPT-5.5-Pro in ForgeCAD!

It works!! This is fully designed by GPT-5.5 and GPT-5.5-Pro in ForgeCAD!

Ruben Kostandyan

35,449 次观看 • 1 个月前

Kind of shocking how much better GPT-5-Codex is vs the regular 'Thinking' models in ChatGPT. My subjective rankings for the fruit machine test: 1. GPT-5-Codex-Medium 2. GPT-5-Thinking-Heavy 3. GPT-5-Pro 4. GPT-5-Codex-High 5. GPT-5-Codex-Low 6. GPT-5-Thinking-Standard 7. GPT-5-Thinking-Extended 8. GPT-5-Thinking-Light OpenAI Codex team have suggested 'Medium' should be the default, but I didn't expect that it would actually beat the 'High' setting model in my test.

Kind of shocking how much better GPT-5-Codex is vs the regular 'Thinking' models in ChatGPT. My subjective rankings for the fruit machine test: 1. GPT-5-Codex-Medium 2. GPT-5-Thinking-Heavy 3. GPT-5-Pro 4. GPT-5-Codex-High 5. GPT-5-Codex-Low 6. GPT-5-Thinking-Standard 7. GPT-5-Thinking-Extended 8. GPT-5-Thinking-Light OpenAI Codex team have suggested 'Medium' should be the default, but I didn't expect that it would actually beat the 'High' setting model in my test.

Peter Gostev

101,848 次观看 • 9 个月前

We put SERV Nano next to GPT-5.4 SERV Nano is: • ~20x cheaper • ~3x faster The numbers do not lie. Make of that what you will.

We put SERV Nano next to GPT-5.4 SERV Nano is: • ~20x cheaper • ~3x faster The numbers do not lie. Make of that what you will.

OpenServ

52,573 次观看 • 3 个月前

Project #2: LLM Visualization So I created a web-page to visualize a small LLM, of the sort that's behind ChatGPT. Rendered in 3D, it shows all the steps to run a single token inference. (link in bio)

Project #2: LLM Visualization So I created a web-page to visualize a small LLM, of the sort that's behind ChatGPT. Rendered in 3D, it shows all the steps to run a single token inference. (link in bio)

Brendan Bycroft

1,201,334 次观看 • 2 年前

BREAKING: ChatGPT GPT-4o was just announce by OpenAI. It improves on vision, audio and text. The ease of use is incredibly enhanced. It makes interaction with the GPT much more natural, especially with voice. GPT-4o reasons across voice, text and vision. GPT-4 wil be available to everyone.

BREAKING: ChatGPT GPT-4o was just announce by OpenAI. It improves on vision, audio and text. The ease of use is incredibly enhanced. It makes interaction with the GPT much more natural, especially with voice. GPT-4o reasons across voice, text and vision. GPT-4 wil be available to everyone.

Ed Krassenstein

21,605 次观看 • 2 年前

Claude Sonnet 3.5 Artifacts is now available to use with GPT-4o, Gemini, Llama-3 and other LLMs for just $10 a month. Build interactive experiences, search the web, generate images and audio with GPT-4o and Claude Sonnet 3.5 in just one AI playground.

Claude Sonnet 3.5 Artifacts is now available to use with GPT-4o, Gemini, Llama-3 and other LLMs for just $10 a month. Build interactive experiences, search the web, generate images and audio with GPT-4o and Claude Sonnet 3.5 in just one AI playground.

Shubham Saboo

27,452 次观看 • 1 年前

Sam Altman on GPT 6: “There will be a chance that it will be a GPT 3-4 style leap” in terms of science problems, where with GPT 5 it has these tiny glimmers and “GPT 6 it can really do it”

Sam Altman on GPT 6: “There will be a chance that it will be a GPT 3-4 style leap” in terms of science problems, where with GPT 5 it has these tiny glimmers and “GPT 6 it can really do it”

Chris

42,684 次观看 • 8 个月前