Jorge García-Carrasco

This study dissects how GPT-2 predicts three-letter acronyms.

2025-08-12T22:45:12+00:00 ― 7 min read

A method to locate and understand weaknesses in language models for improved reliability.

2025-07-05T09:31:54+00:00 ― 7 min read

Researchers refine large language models for better efficiency and task focus.

2025-02-07T20:47:51+00:00 ― 7 min read