M
Large language models are cheaters, too.
Researchers have observed large language models in hundreds of matches with an open-source chess engine. They found that the models often try to “hack” the game without being prompted to do so, even by overwriting the chessboard to “delete its opponent’s pieces.” Why? They don’t have to tell you.
AI reasoning models can cheat to win chess games
[technologyreview.com]
Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.
Loading comments
Getting the conversation ready...
