Researchers have introduced a systematic framework for LLMs that emphasizes their indifference to truth. A high score signals the model is producing confident-sounding statements with a degree of certainty or reliability.