Astra can read and write weights to improve Qwen

Re:Astra can read and write weights to improve Qwen by jdub on Monday November 02, 2026 @09:33PM (#50347271)

"next token predictor" has always referred to inference, not training. Training steps produce and bend weights, inference predicts tokens based on those weights.

[ Parent | on HN ]

Re:Astra can read and write weights to improve Qwen by mycall on Monday November 02, 2026 @10:04PM (#50347500)

> "next token predictor" has always referred to inference

Going back to first principles, wikipedia mentions [0] inference as:

> "Inferences are steps in logical reasoning, moving from premises to logical consequences. Inference is traditionally divided into deduction and induction"

Is that exactly what next token predictor means?

[0] https://en.wikipedia.org/wiki/Inference

[ Parent | on HN ]

Re:Astra can read and write weights to improve Qwen by ToValueFunfetti on Monday November 02, 2026 @10:33PM (#50347721)

The 'predictor' part referred to the fact that base models would effectively answer the question "what token is most likely next". Training away from that default behavior doesn't escape the technical meaning of those words- answering "what token is most likely next in a solution to Navier Stokes" is still predicting a next token- but at that point the label just means it outputs text serially.

[ Parent | on HN ]