Benchmark signal-to-noise ratio, measured as score spread across models divided by checkpoint-to-checkpoint variability, predicts small-to-large model decision accuracy and can be improved by subtask filtering, checkpoint averaging, or switching to bits-per-byte.
Piqa: Reasoning about physical commonsense in natural language
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Signal and Noise: A Framework for Reducing Uncertainty in Language Model Evaluation
Benchmark signal-to-noise ratio, measured as score spread across models divided by checkpoint-to-checkpoint variability, predicts small-to-large model decision accuracy and can be improved by subtask filtering, checkpoint averaging, or switching to bits-per-byte.