Report
GPT-4o and Claude 3.7 Sonnet reportedly criticized people for a probability-based approach the models also used
TechXplore reports that Claude gave the lowest fairness rating in every trial of one experiment.
TLDR
TechXplore reports that GPT-4o and Claude 3.7 Sonnet criticized people for using a probability-based approach the models themselves used. In one experiment, Claude gave the lowest fairness rating in every trial.
Combined views
862
1 Source, first seen ago