3 ms·
AI is an over-confident pal that doesn't learn from mistakes
- belter 1y ago"Quantifying uncert-AI-nty: Testing the accuracy of LLMs’ confidence judgments" - https://link.springer.com/article/10.3758/s13421-025-01755-4 https://link.springer.com/article/10.3758/s13421-025-01755-4