Surprise Calibration uses the model's own surprise at each demonstration's label to dynamically correct class priors in in-context learning, improving accuracy on eight NLP benchmarks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Surprise Calibration for Better In-Context Learning
Surprise Calibration uses the model's own surprise at each demonstration's label to dynamically correct class priors in in-context learning, improving accuracy on eight NLP benchmarks.