Comment by NomDePlum

Really nice simple to understand article. I do wonder if the capabilities demonstrated here are specific to the LLM in use in the article or would most models display similar behaviour?