@poetaster yeah not really what people associate with LLMs, and likely they work best with highly specific models, not generic LLMs