factual

LLMs Generalize Poorly Out of Distribution Despite Vast Data

Large models require enormous amounts of training data yet generalize poorly, especially to out-of-distribution cases not present in their training, whereas children with vastly less data generalize very well; LLMs notably lack the capacities to do real-world experiments and build abstract causal models that support generalization.

factualpending

Speaker

Alison Gopnik

Evidence Quote

“The large models need enormous amounts of data, and they're not very good at generalizing, especially to what's called out of distribution cases, cases that weren't there in their training. The kids with much less data, are very good at doing that”

Source

308 | Alison Gopnik on Children, AI, and Modes of Thinking— Sean Carroll's Mindscape
Created: 6/13/2026, 12:24:35 AM

Your notes

Loading notes...