Article

Large Language Models in the Labyrinth: Possibility Spaces and Moral Constraints

2023-10-17

Abstract excerpt

<p>To think about possibilities requires that we navigate an almost unimaginably large and labyrinthine space of what could be. Heuristics can make this problem tractable for the human mind, but we understand little about the underlying structures these heuristics operate on. This paper shows how large language models (LLMs, and GPT-4 in particular) can provide a new window onto the challenges that both humans and...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: a19b6944-948e-5d8f-9724-c6b14f32344a
DOI: 10.31234/osf.io/zvk7b

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

Large Language Models in the Labyrinth: Possibility Spaces and Moral ConstraintsDOI 10.31234/osf.io/zvk7bRead

LLM-generated possibilities increase blame attributionSemantic proximity Read

How humans process language (they think) is machine generatedSemantic proximity Read

Nonlinear Transformation of Probabilities by Large Language ModelsSemantic proximity Read

Form Theory: The Conceptual Architecture of Human ThoughtSemantic proximity Read

The Moral Turing Test: Evaluating Human-LLM Alignment in Moral Decision-MakingSemantic proximity Read

When Do LLMs Say “Some”? An Investigation of Scalar Implicature and Politeness Mitigation in Large Language ModelsSemantic proximity Read

Similar failures of consideration arise in human and machine planningSemantic proximity Read

Generating options and choosing between them rely on distinct forms of value representationSemantic proximity Read

Generating options and choosing between them rely on distinct forms of value representationSemantic proximity Read

Do LLMs have core beliefs?Semantic proximity Read

In Praise of Folly: Flexible Goals and Human CognitionSemantic proximity Read

Rethink Your Mental Model in the Age of Generative AI: A Triadic Framework for Human-AI CollaborationSemantic proximity Read