Out-of-Context Reasoning (OOCR) in LLMs: A Short Primer and Reading List

·LessWrong··

Out-of-context reasoning (OOCR) is a concept relevant to LLM generalization and AI alignment. Also available as a PDF. Contents What is OOCR? Examples Papers Videos What is out-of-context reasoning for LLMs? It's when an LLM reaches a conclusion that requires non-trivial reasoning but the reasoning is not present in the context window. The reasoning could instead take place in the forward pass or during the training process. The name ("out-of-context reasoning") is chosen to contrast with in-con...

Read full article →

Related Articles

OpenWrt One – Open Hardware Router
peter_d_sherman · Hacker News · 11h ago
GLM 5.2 and the coming AI margin collapse
martinald · Hacker News · 9h ago
What Emily Bender meant by "stochastic parrots"
digital55 · Hacker News · 14h ago
Nintendo announces new product revisions in Europe with replaceable batteries
akyuu · Hacker News · 16h ago
A global workspace in language models
in-silico · Hacker News · 11h ago