Code in pre-training data improves LLM performance at non-coding tasks [VentureBeat]

View Article on VentureBeat