Abstract: Estimation of a model’s confidence on its outputs is critical for Conversational AI systems based on large language models (LLMs), especially for reducing hallucination and preventing ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback