Algorithmic Reasoning Evolution
Tim and Hattie discuss the evolution of algorithmic reasoning in language models, highlighting the shift towards giving models programs to extrapolate solutions independently. They delve into the concept of inductive algorithmic reasoning, exploring the spectrum from explicit step-by-step algorithms to more vaguely described processes.In this clip
From this podcast
Machine Learning Street Talk (MLST)
#91 - HATTIE ZHOU - Teaching Algorithmic Reasoning via In-context Learning #NeurIPS
Related Questions