Astra can read and write weights to improve Qwen

Re:Astra can read and write weights to improve Qwen by BrokenCogs on Monday November 02, 2026 @09:30PM (#50347251)

I believe the authors are wildly overstating the conclusions in their abstract.

Did they test modifying the weights of a non-public model?

Did they test modifying the weights of a non-LLM (and non-public) model?

If not, we can only conclude that GPT-6 Astra is good at reading and writing weights of Qwen (a public LLM), not just any neutral network.

[ Parent | on HN ]

Re:Astra can read and write weights to improve Qwen by leumassuehtam on Monday November 02, 2026 @09:36PM (#50347290)

The result is still surprising even if the result is Astra can edit and write weights of a publicly known model. It means that models can, at least in principle, be trained on a task that it wasn't considered possible before.

[ Parent | on HN ]