DECAF decomposes paired perturbation responses into endpoint-aligned evidence, endpoint-opposed contradiction, and endpoint-null fragility, and its components track independently measured model behavior better than raw magnitude on controlled and natural-image benchmarks.
The Annals of Mathematical Statistics , volume=
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
-
Decomposition of Evidence, Contradiction, and Fragility in Perturbation Responses
DECAF decomposes paired perturbation responses into endpoint-aligned evidence, endpoint-opposed contradiction, and endpoint-null fragility, and its components track independently measured model behavior better than raw magnitude on controlled and natural-image benchmarks.