This article was originally published in ai/p/llm-codenames-competition” rel=”noopener ugc nofollow” target=”_blank”>Artistic fish intelligence.
LLMs are good at many things and one of them is playing games. People have used LLMs to play all kinds of games, such as minecraft, Chess, murder mystery games, Werewolfand the NYT connection puzzle. (For a more complete list, you can consult this survey.)
Most of the examples above show LLMs playing against themselves or against humans. But how well do LLMs play against other LLM?
In this article, I show the results of three different LLMs competing against each other in the popular board game Codenames, which challenges players to find patterns between seemingly unrelated words.
For those who don't know him, Code names is a board game created by Vladimir Chvátil. The game pits two teams (usually red and blue) against each other.
Each team has a spymaster who gives single-word clues meant to point out multiple words on a 5×5 word board. The other players on the team must guess the words of their team, avoiding the words that belong to the opposing team. Only the…
