Skip to main content
Search

Roksana Goworek

1 related articles

Subscribe to updates
  1. arXiv

    Frontier LLMs can write small network weights outright: replication holds, but reading weights mostly reads the input layer

    A three-day replication tested whether frontier LLMs can read and write the weights of small neural networks directly. GPT-6 Astra's Breakout networks scored 514, 550 and 307 across three runs, in line with the originally reported 517, and did just as well on two Breakout variants that exist nowhere else; yet when two models independently decompiled Astra's Breakout network into Python, both programs chose the same action as the network in all 49,189 game states tested, indicating that writing weights amounts to compiling a program into weights while reading weights mostly recovers the input-layer layout.

You’re all caught up