The Machine's Internal Clock: Do LLMs Share Human Temporal Illusions?

Name: The Machine's Internal Clock: Do LLMs Share Human Temporal Illusions?
Uploaded: 2026-10-22T03:55:50.969984+00:00
Duration: 327 s
Description: Researchers tested if Large Language Models (LLMs) exhibit human-like temporal illusions using written narratives. A benchmark of 6,684 scenarios spanning five illusions was used. Human readers showed reliable effects for only two illusions, while LLMs reproduced patterns for four. This suggests model alignment stems from retrieving literature rather than genuine human temporal biases.

Researchers tested if Large Language Models (LLMs) exhibit human-like temporal illusions using written narratives. A benchmark of 6,684 scenarios spanning five illusions was used. Human readers showed reliable effects for only two illusions, while LLMs reproduced patterns for four. This suggests model alignment stems from retrieving literature rather than genuine human temporal biases.

Open the full episode