Across 140k generations, wrapper-only formatting changes accuracy by over 30× across models, largely via parseability failures, so single-wrapper scores are statistically fragile.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Format Sensitivity Index: Token-Controlled Prompt Wrapper Robustness and Schema Compliance in LLM Benchmarking
Across 140k generations, wrapper-only formatting changes accuracy by over 30× across models, largely via parseability failures, so single-wrapper scores are statistically fragile.