4 ms·
The new idea is inference-time scaling, as seen in o1 (and o3 and Qwen's QwQ and DeepSeek's DeepSeek-R1-Lite-Preview and Google's gemini-2.0-flash-thinking-exp)
by simonw 2y ago
The new idea is inference-time scaling, as seen in o1 (and o3 and Qwen's QwQ and DeepSeek's DeepSeek-R1-Lite-Preview and Google's gemini-2.0-flash-thinking-exp).
I suggest reading these two pieces about that:
- https://www.aisnakeoil.com/p/is-ai-progress-slowing-down https://www.aisnakeoil.com/p/is-ai-progress-slowing-down - best explanation I've seen of inference scaling anywhere
- https://arcprize.org/blog/oai-o3-pub-breakthrough https://arcprize.org/blog/oai-o3-pub-breakthrough - François Chollet's deep dive into o3
I've been tracking it on this tag on my blog: https://simonwillison.net/tags/inference-scaling/ https://simonwillison.net/tags/inference-scaling/
- exhaze 2y agoI think the wildest thing is actually Meta’s latest paper where they show a method for LLMs reasoning not in English, but in latent space https://arxiv.org/pdf/2412.06769 https://arxiv.org/pdf/2412.06769 I’ve done research myself adjacent to this (mapping parts of a latent space onto a manifold), but this is a bit eerie, even to me.
- asadalt 2y agokinda how we do it. language is just an io interface(but also neural obv) on top of our reasoning engine.
- oceanparkway 2y agoIt’s not just a protocol buffer for concepts though (weak wharf Sapir, lakoff’s ubiquitous metaphors). Language itself is also a concept layer and plasticity and concept development is bidirectional. But (I’m not very versed in the language here re ‘latent space’) I would imagine the forward pass through layers converges towards near-token-matches before output, so you have very similar reason to token/language reasoning even in latent/conceptual reasoning? Like the neurons that nearly only respond to a single token for ex.
- ynniv 2y agoIs it "eerie"? LeCun has been talking about it for some time, and may also be OpenAI's rumored q-star, mentioned shortly after Noam Brown (diplomacybot) joining OpenAI. You can't hill climb tokens, but you can climb manifolds.
- Y_Y 2y ago> You can't hill climb tokens, but you can climb manifolds. Could you explain this a bit please?
- sebzim4500 2y agoI imagine he means that when you reason in latent space the final answer is a smooth function of the parameters, which means you can use gradient descent to directly optimize the model to produce a desired final output without knowing the correct reasoning steps to get there. When you reason in token space (like everyone is doing now) you are executing nonlinear functions when you sample after each token, so you have to use some kind of reinforcement learning algorithm to learn the weights.
- Y_Y 2y agoI think there's a subtlety here about what makes (e.g. English) tokens different to points in latent space. Everything is still differentiable (at least in the ML sense) until you do random sampling. Even then you can exclude the sampling when calculating the gradient (or is this equivalent to the "manifold"?). I don't see a priori why it would be better or worse to reason with the "superposition" of arguments in the pre-sampling phase rather than concrete realizations of those arguments found only after choosing the token. It may well be a contingent rather than necessary fact.
- ynniv 2y agoLinks to Yan: Title: "Objective Driven AI: Towards Machines that can Learn, Reason, and Plan" Lytle Lecture Page: https://ece.uw.edu/news-events/lytle-lecture-series/ https://ece.uw.edu/news-events/lytle-lecture-series/ Slides: https://drive.google.com/file/d/1e6EtQPQMCreP3pwi5E9kKRsVs2NbWPrY/view?usp=drivesdk https://drive.google.com/file/d/1e6EtQPQMCreP3pwi5E9kKRsVs2N... Video: https://youtu.be/d_bdU3LsLzE?si=UeLf0MhMzjXcSCAb https://youtu.be/d_bdU3LsLzE?si=UeLf0MhMzjXcSCAb
- danielmarkbruce 2y agoIt's just concept space. The entire LLM works in this space once the embedding layer is done. It's not really that novel at all.
- ttul 2y agoThis was my thought. Literally everything inside a neural network is a “latent space”. Straight from the embeddings that you use to map categorical features in the first layer. Latent space is where the magic literally happens.
- madethisnow 2y agoCompletely agree. Have you see this? https://sakana.ai/asal/ https://sakana.ai/asal/
- mountainriver 2y agoThere are lots of papers that do this
- rhubarbtree 2y agoSeems a standard approach of AI research is to “move X into the latent space” where X is some useful function (eg diffusion) previously done in the “data” or “artefact” space. So seems very pedestrian not wild to make that step.