In this work, we introduce a new human evaluation of #LLMs called Representation Alignment. Representation Alignment looks at the similarity between how LLMs use conceptual representations and how humans interpret them.

Or maybe your dreams as a janitor will lead you to be the captain of your very own garbage ship!

The LARA Lab has a new paper out looking at the alignment (or lack there of) of #LLMs like #GPT and #LLaMA with human perceptions of emotions. Turns out the representation of the emotion matters!

Shout out to @namer.bsky.social and Asha Kumar for their amazing work!

arxiv.org/abs/2503.11881

Since when has our mascot basically become a shrine? It's kind of cute. #umbc

Unlike paying someone to do your assignment or copying a friend's, students don't seem to know how well AI is actually doing the assignment. We're really going to see a paradigm shift in teaching where we teaching higher-level skills (analyzing and evaluating) far earlier on than we used to.

Students from my Interactive Fiction and Text Generation class presented their final projects today, and it was amazing! I opened it up to the department, and faculty & students got to see all the fun AI work people did this semester. There were so many cool research projects and demos.