two copies of gpt-3.5 (one playing as the oracle, and another as the guesser) performs poorly on the game of 20 Questions (68/1823).

Follow the full discussion on Reddit.
I put two copies of gpt-3.5 as partners, one plays the role of the oracle that answers yes/no questions, the other as the role of a guesser that asks yes/no questions. I want to see if gpt-3.5 would perform well on this "dynamic" task -- i.e. rather than a fixed test set with 1 good answer, 20 questions can be played into many paths, depending on the questions being asked.

Comments

There's unfortunately not much to read here yet...

Discover the Best of Machine Learning.

Ever having issues keeping up with everything that's going on in Machine Learning? That's where we help. We're sending out a weekly digest, highlighting the Best of Machine Learning.

Join over 900 Machine Learning Engineers receiving our weekly digest.

Best of Machine LearningBest of Machine Learning

Discover the best guides, books, papers and news in Machine Learning, once per week.

Twitter