Claude Mythos and the Case for Looped Transformers
中文摘要
探讨循环Transformer架构及其在测试时缩放和内部潜在推理方面的作用。
English Summary
Exploring looped Transformers, focusing on test-time scaling and latent reasoning within the model.
Original Excerpt
From test-time scaling to latent reasoning inside the model Continue reading on Medium »