The inaugural Conference on Language Modeling is a new academic venue that aims to build a community of researchers focused on understanding, improving, and critiquing the development of language modeling technology.
This year, four papers received COLM’s Outstanding Paper Award. Among those is a paper on “Auxiliary task demands mask the capabilities of smaller language models.” Co-authored by Jennifer Hu (a research fellow at Harvard’s Kempner Institute who will join our cognitive science faculty next year) and Michael C. Frank (faculty in the Sandford Department of Psychology), this paper demonstrates how researchers’ design choices matter when it comes to the performance capacities of smaller language models.