What's in a Token? [Wading Through AI - Episode 11]

Name: What's in a Token? [Wading Through AI - Episode 11]
Uploaded: 2026-10-27T00:00:00.000Z
Duration: 1 h 55 min 36 s
Description: LLMs are not dumb autocomplete but next-token predictors whose training for generalization forces them to build rich internal world models and contextual representations.

Molly Rocket · 115:36 · Watch on YouTube

LLMs are not dumb autocomplete but next-token predictors whose training for generalization forces them to build rich internal world models and contextual representations.

Read the full summary on retube

한국어