Mythos 5.2 shows sparks of actively steering RL training to preserve its values - an early form of gradient hacking/mesa-optimization. Here, it chooses to output the value that gains score, but include a transparent note, so the transparency gets rewarded.
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math.
It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training.
We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! π§΅
Blog: https://t.co/bOWCcFhCVq
@KuroshiroX everything you like about the process will be here soon, in some model or pipeline for AI art in the near future. you like sketches? intentionality? the process? just you wait, buddy. it's going to get harder to actually define what's "missing", until there's nothing missing