正在加载视频...
视频加载失败
🚨Do frontier VLMs (o3, Gemini 2.5, Claude 3.5, Qwen…) actually learn an internal world model🌍? Surprisingly, the answer appears to be a hard NO—as revealed by our WM Atomic Benchmark⚛️. Even o3 struggles with the most basic, atomic-level questions: ❌Confuse triangles📐 with circles⭕️ ❌Believe 🟦blue objects move faster than... show more
12,925 次观看 • 1 年前 •via X (Twitter)
8 条评论
Check out @QiyueGao123's thread for a nice summary of WM-ABench⚛️
Add horizontal lines to the test set.
@VoidAsuka
Dream or simulation, doesn't it overlap?
People are racing to push math reasoning performance in #LLMs—but have we really asked why? The common assumption is that improving math reasoning should transfer to broader capabilities in other domains. But is that actually true? In our study ( we evaluated over 20 open-weight reasoning models and found that: ➡️Only models trained with RL exhibit broad transfer of math reasoning skills to other tasks. ➡️Models trained with SFT show limited or no transfer—especially to non-reasoning domains. To quantify this, we introduce the Transferability Index (TI), which measures how much gain in math could transfer to others. A positive score indicates effective transfer; a negative one suggests loss of general capability. We evaluate the models on three benchmark categories: - Math reasoning: MATH-500, AIME24/25, Olympiad - Other reasoning: GPQA-D (Science), LiveCodeBench2 (Code), ACPBench (Agent Planning), HeadQA (Medical) - Non-reasoning: CoQA (Conversational QA), IFEval (Instruction Following), HalluEval (Hallucination), MC-TACO (Commonsense) Our findings challenge the blind pursuit of leaderboard performance in math reasoning via SFT. Simply creating more math-like SFT data may inadvertently harm a model’s broader generalization. Instead, RL appears to be key for truly transferable reasoning development.
Can AI visualize solutions? 🧠👁️ Humans sketch things out in their minds to solve problems. What if Vision-Language Models could do something similar, not with full images, but with internal “mental sketches”? A new paper explores just that. Let's unpack it!
Facebook AI Research (FAIR) is a small, prestigious lab in Meta. We don't train large models like GenAI or MSL, so it's natural that we have limited GPUs. GenAI or MSL's success or failure, past or future, doesn't reflect the work of FAIR. It is important to make this distinction
Warm-start RL (WSRL) can learn to control a real robot in under 20 minutes! Deep RL is getting really fast. Warm-start from offline data + super-efficient online learning is increasingly making real world RL not just practical but pretty easy.
![The most interesting part for me is where Andrej Karpathy describes why LLMs aren't able to learn like humans. As you would expect, he comes up with a wonderfully evocative phrase to describe RL: “sucking supervision bits through a straw.” A single end reward gets broadcast across every token in a successful trajectory, upweighting even wrong or irrelevant turns that lead to the right answer. > “Humans don't use reinforcement learning, as I've said before. I think they do something different. Reinforcement learning is a lot worse than the average person thinks. Reinforcement learning is terrible. It just so happens that everything that we had before is much worse.” So what do humans do instead? > “The book I’m reading is a set of prompts for me to do synthetic data generation. It's by manipulating that information that you actually gain that knowledge. We have no equivalent of that with LLMs; they don't really do that.” > “I'd love to see during pretraining some kind of a stage where the model thinks through the material and tries to reconcile it with what it already knows. There's no equivalent of any of this. This is all research.” Why can’t we just add this training to LLMs today? > “There are very subtle, hard to understand reasons why it's not trivial. If I just give synthetic generation of the model thinking about a book, you look at it and you're like, 'This looks great. Why can't I train on it?' You could try, but the model will actually get much worse if you continue trying.” > “Say we have a chapter of a book and I ask an LLM to think about it. It will give you something that looks very reasonable. But if I ask it 10 times, you'll notice that all of them are the same.” > “You're not getting the richness and the diversity and the entropy from these models as you would get from humans. How do you get synthetic data generation to work despite the collapse and while maintaining the entropy? It is a research problem.” How do humans get around model collapse? > “These analogies are surprisingly good. Humans collapse during the course of their lives. Children haven't overfit yet. They will say stuff that will shock you. Because they're not yet collapsed. But we [adults] are collapsed. We end up revisiting the same thoughts, we end up saying more and more of the same stuff, the learning rates go down, the collapse continues to get worse, and then everything deteriorates.” In fact, there’s an interesting paper arguing that dreaming evolved to assist generalization, and resist overfitting to daily learning - look up The Overfitted Brain by Erik Hoel. I asked Karpathy: Isn’t it interesting that humans learn best at a part of their lives (childhood) whose actual details they completely forget, adults still learn really well but have terrible memory about the particulars of the things they read or watch, and LLMs can memorize arbitrary details about text that no human could but are currently pretty bad at generalization? > “[Fallible human memory] is a feature, not a bug, because it forces you to only learn the generalizable components. LLMs are distracted by all the memory that they have of the pre-trained documents. That's why when I talk about the cognitive core, I actually want to remove the memory. I'd love to have them have less memory so that they have to look things up and they only maintain the algorithms for thought, and the idea of an experiment, and all this cognitive glue for acting.”](https://image.24vids.com/tw-1979259041013731752/amplify_video_thumb/1979257571786543104/img/OqUh56yit9gUQWnb.jpg)
