I wanted to test this claim with SAT problems. Why SAT? Because solving SAT problems require applying very few rules consistently. The principle stays the same even if you have millions of variables or just a couple. So if you know how to reason properly any SAT instances is solvable given enough time. Also, it's easy to generate completely random SAT problems that make it less likely for LLM to solve the problem based on pure pattern recognition. Therefore, I think it is a good problem type to test whether LLMs can generalize basic rules beyond their training data.
Раскрыты подробности похищения ребенка в Смоленске09:27，更多细节参见51吃瓜
Credit: Hisense，更多细节参见谷歌
Названо необходимое для чистого воздуха количество растений в доме14:53
当然了，摸着良心说，gpt-oss-120B 不是 OpenAI 的旗舰产品，而是其开放权重的中端线。而且它采用 MoE 架构，标称 120B 参数，但每个 token 实际只激活约 5.1B 参数——所以参数量的对比，在工程层面并不像字面数字那么悬殊。