Echoes of Intelligence.

This article focuses on the aspect of textual interpretation in a large language model (LLM). The author cites works by a variety of authors to explain how an LLM works, including Angela Friederici’s two types of semantic representations, Umberto Eco’s works on semantic interpretation and reader’s codes to obtain meaning from text, and Jean-François Lyotard’s definitions for utterances defined by rules. Topics also discussed include the role of the LLM reader and how to build one.