LATIDIA · Agentes

Qwen 3.8 27B en mi portátil también puede interpretar los pesos de los modelos

El documento de ayer fue Sparks of Weight-Space Programming: los LLM pueden leer y escribir de forma nativa los pesos de las redes neuronales de Rikiya Takehi, Jayden Teoh, Alex Zhang y Omar Khattab. El artículo (OK, los autores clasifican es

La noticia

Yesterday's hot paper was Sparks of Weight-Space Programming: LLMs can natively read & write neural network weights by Rikiya Takehi, Jayden Teoh, Alex Zhang, and Omar Khattab. The paper (OK, the authors classify is as a "blog post", but it's paper-shaped) found that frontier models, in particular GPT-6 Astra, can both interpret and modify the raw weights of a machine learning model - literally a text description of the model containing a bunch of floating point numbers. I've been playing with one of the examples from the paper and I'm delighted to share that Qwen 3.8 27B running on my laptop can do this as well! Weight-Space Programming Here's the