Rikiya Takehi:

Researchers say LLMs can be good at natively reading and writing neural network weights; GPT-6 Astra edited Qwen3.5-0.8B's weights to make it better at math  —  TL;DR … If you were asked to handcraft a neural network that plays Go, what would you do?  If you were handed raw neural network weights …

Rikiya Takehi:

Researchers say LLMs can be good at natively reading and writing neural network weights; GPT-6 Astra edited Qwen3.5-0.8B's weights to make it better at math  —  TL;DR … If you were asked to handcraft a neural network that plays Go, what would you do?  If you were handed raw neural network weights …

Source: TechMeme
Source Link: https://www.techmeme.com/261102/p36#a261102p36