Ask HN: Why LLM bad at memory/follow instruction?1 point·ihsanf··1 commentwhy if we test different models to role play a character, some of them can do it really good with shorter system prompt but the rest are still bad even with longer prompt?SaveView on HN