1 paper
Sravanthi Machcha, Sushrita Yerra, Sahil Gupta +4
Current evaluation of large language models (LLMs) overwhelmingly prioritizes accuracy; however, in real-world and safety-critical applications, the ability to abstain when uncerta…