☆ 4.3 Article

Generative AI Are More Truth-Biased Than Humans: A Replication and Extension of Core Truth-Default Theory Principles

JOURNAL OF LANGUAGE AND SOCIAL PSYCHOLOGY (2023)

Related references

Note: Only part of the references are listed.

Article Psychology, Biological

Emergent analogical reasoning in large language models

Taylor Webb et al.

Summary: Webb et al. demonstrate that new AI language models, such as GPT-3, can solve analogical reasoning problems at a human-like level. The emergence of large language models has sparked a debate on whether these generic models can exhibit human cognitive abilities with sufficient training data. Particularly interesting is their ability to reason zero-shot, without direct training, which is closely tied to analogical reasoning in human cognition. Through a comparison with human reasoners, GPT-3 displayed a surprisingly strong capacity for abstract pattern induction, even surpassing human capabilities, and preliminary tests of GPT-4 showed even better performance. This signifies that large language models like GPT-3 have acquired an emergent ability to find zero-shot solutions to a wide range of analogy problems.

NATURE HUMAN BEHAVIOUR (2023)

Add to Collection

Article Criminology & Penology