bandit

Experiment 28

12 minute read

I tested four bandit algorithms on news recommendation to figure out if contextual learning could teach LLMs to adapt their prompting style to individual use...