LLMs
Meta Deploys LLM-Generated Tests That Catch Bugs Before Code Lands
Meta has moved its Catching Just-in-Time Tests (JiTTests) from research prototype to production infrastructure. According to Meta, the system uses large language models…
Google Research Gives LLMs a Sleep Phase to Retain Knowledge
Google Research has introduced a sleep framework for continual learning, in which a large language model alternates between active use and an offline phase that…
Apple study: efficient LLM steering methods often come at a steep cost to fluency
A new Apple machine learning paper finds that efficient methods for steering large language models frequently achieve their goal at a steep cost to fluency. The study…