@percyliang : Most leaderboards just give you scores, leaving one wondering: what does 76.8% mean? In HELM, we are committed to full transparency, meaning clicking on a score will reveal the full set of instances, and you can even inspect the exact prompt (which we know makes a big • TwiCopy

Percy Liang

@percyliang

+ Follow

Associate Professor in computer science @Stanford @StanfordHAI @StanfordCRFM @StanfordAILab @stanfordnlp | cofounder @togethercompute | Pianist

ID: 86481377

linkhttps://cs.stanford.edu/~pliang/ calendar_today31-10-2009 07:26:37

905 Tweet

55,55K Followers

412 Following

Percy Liang

@percyliang

5 months ago

Most leaderboards just give you scores, leaving one wondering: what does 76.8% mean? In HELM, we are committed to full transparency, meaning clicking on a score will reveal the full set of instances, and you can even inspect the exact prompt (which we know makes a big