Fun Libs
Humor Is So Subjective
A classroom experiment in computational humor, based on Hossain et al. (2017). Fill in the same story twice using two different AI word-selection approaches, then rate the results.
ERA 2 — Statistical LM (c.2010)
ERA 7 — Neural LM + RLHF (2026)
Choose a story
How it works: Click a blank box to see word suggestions from both eras side by side. Your
first word choice locks your model for the entire run. Fill all blanks, view your completed story, then try
the other approach and run it again. At the end, rate both stories and discuss what the difference tells you
about AI.
Click a blank to see word suggestions.
Your first choice locks your model for this run.