Ai Interpretability News Today

Researchers find large language models use a simple mechanism to retrieve stored knowledge when they respond to a user prompt. These mechanisms can be leveraged to see what the model knows about different subjects and possibly to correct false information it has stored.

“Unpredictability may be something we look for in intelligence, and if so, then by definition, a true intelligence will be unpredictable and therefore uninterpretable,” says Toyama.