A case for LLMs as Self-predictors — LessWrong

lesswrong.com · 4,753 words · saved by 1 readers

Written as part of the MATS 9.1 extension program, mentored by Richard Ngo. Additional thanks to Maria Kostylew for helpful draft feedback. …

x A case for LLMs as Self-predictors — LessWrong Agent Foundations Game Theory Language Models (LLMs) MATS Program AI Frontpage 32 A case for LLMs as Self-predictors by Ashe Vazquez Nuñez 5th Jul 2026 12 min read 6 32 Written as part of the MATS 9.1 extension program, mentored by Richard Ng o. Additional thanks to Maria Kostylew for helpful draft feedback. Introduction This post advocates a perspective of LLMs as seeking to minimise prediction error with respect to their world models. We can moreover interpret token outputs and their scaffolded consequences as actions that close a control loop

A case for LLMs as Self-predictors — LessWrong

related reading