In Dialogue with Intelligence: Rethinking Large Language Models as Collective Knowledge

TL;DR

This paper reinterprets Large Language Models as Collective Knowledge, emphasizing their emergent dialogue-based intelligence, dynamic behavior, and potential as a neuroscience research object through human-model interaction.

Contribution

It introduces the concept of Collective Knowledge in LLMs, explores response modes and co-augmentation, and proposes using LLMs as accessible models for neuroscience studies.

Findings

LLMs exhibit distinct response modes linked to model subnetworks.

CK has no persistent internal state, its behavior is shaped by users and fine-tuning.

The human-CK loop can serve as an experimental platform for neuroscience.

Abstract

Large Language Models (LLMs) can be understood as Collective Knowledge (CK): a condensation of human cultural and technical output, whose apparent intelligence emerges in dialogue. This perspective article, drawing on extended interaction with ChatGPT-4, postulates differential response modes that plausibly trace their origin to distinct model subnetworks. It argues that CK has no persistent internal state or ``spine'': it drifts, it complies, and its behaviour is shaped by the user and by fine-tuning. It develops the notion of co-augmentation, in which human judgement and CK's representational reach jointly produce forms of analysis that neither could generate alone. Finally, it suggests that CK offers a tractable object for neuroscience: unlike biological brains, these systems expose their architecture, training history, and activation dynamics, making the human--CK loop itself an experimental target.