Uncertainty-aware fine-tuning with a decision-theory-based loss produces better-calibrated uncertainty estimates than standard training on free-form QA tasks.
Accuracy-rejection curves (arcs) for comparing classification methods with a reject option
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Enhancing Trust in Large Language Models via Uncertainty-Calibrated Fine-Tuning
Uncertainty-aware fine-tuning with a decision-theory-based loss produces better-calibrated uncertainty estimates than standard training on free-form QA tasks.