A recipe that maps stakeholder risk tolerance to a confidence level and then uses statistical methods, with conformal prediction performing best, to set thresholds on LLM evaluation metrics.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ML 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
How to Choose a Threshold for an Evaluation Metric for Large Language Models
A recipe that maps stakeholder risk tolerance to a confidence level and then uses statistical methods, with conformal prediction performing best, to set thresholds on LLM evaluation metrics.