With the rapid progress of models like GPT-4, Claude, and Llama 2, large language models have demonstrated remarkable capability across a wide range of tasks. Yet they remain fundamentally generic — they give the same answer to the same question regardless of who is asking.
