Will's Blog: Researchers detail Centaur, a modified version of Llama trained on 10M+ responses from 160 psychological experiments, in an attempt to mimic human cognition (Carl Zimmer/New York Times)

3 July 2025

Researchers detail Centaur, a modified version of Llama trained on 10M+ responses from 160 psychological experiments, in an attempt to mimic human cognition (Carl Zimmer/New York Times)

Carl Zimmer / New York Times:
Researchers detail Centaur, a modified version of Llama trained on 10M+ responses from 160 psychological experiments, in an attempt to mimic human cognition — To better understand human cognition, scientists trained a large language model on 10 million psychology experiment questions.

Posted from: this blog via Microsoft Power Automate.