Large Language Models’ Emergent Abilities Are a Mirage
A new study suggests that sudden jumps in LLMs’ abilities are neither surprising nor unpredictable, but are actually the consequence of how we measure ability in AI.
A new study suggests that sudden jumps in LLMs’ abilities are neither surprising nor unpredictable, but are actually the consequence of how we measure ability in AI.
By rewarding computers that combined different approaches to solve chess puzzles, Google created an enhanced AI that could defeat its existing champion, AlphaZero.