Skip to content

Documentary state: DRAFT_UNACCEPTED. This draft has NOT been editorially admitted. Mathematical claims require independent checking.

OAI-26 family 132: a claimed superquadratic sensitivity separation

Editorial status: DRAFT_UNACCEPTED. NO_CERTIFICATION. This source-pinned documentary account describes the publisher's claimed construction and does not independently verify its proof.

The question in plain language

How many input bits can meaningfully affect the output of a Boolean function? Sensitivity measures how many individual bits can flip the output at a given input. Block sensitivity instead counts disjoint groups of bits, each of which can flip the output when the entire group is changed.

The quadratic sensitivity conjecture proposed that block sensitivity should never grow faster than the square of sensitivity, up to constant factors. Family 132 claims a family of counterexamples with a strictly larger exponent.

A small example

Take the AND of three bits. At input 111, changing any one bit changes the output, so three individual bits are sensitive. At input 000, no single bit changes the result, but flipping all three at once does. More complicated Boolean functions can make disjoint blocks influential even while individual flips are relatively constrained.

For a total Boolean function \(f\), write \(s(f)\) for the maximum number of sensitive coordinates at any input, and \(\mathrm{bs}(f)\) for the maximum number of pairwise-disjoint sensitive blocks at any input. Both maxima range over all binary inputs, and their maximizing inputs need not coincide.

What is known

The protected Family 132 catalogue identifies one manuscript, A superquadratic separation between sensitivity and block sensitivity.

The pinned publisher overview claims a family of total Boolean functions and a fixed constant \(\alpha>2\) satisfying

\[ \mathrm{bs}(f)\ge s(f)^\alpha \]

along the constructed family. A meaningful asymptotic separation requires an unbounded sequence of sensitivities; the theorem's exact quantifiers must be read from the manuscript before adopting this inequality as a proven counterexample.

The claimed consequence is a refutation of the quadratic strengthening of the Sensitivity Conjecture. It does not refute the established polynomial relation between sensitivity and block sensitivity. Nor should it be conflated with statements about partial Boolean functions, where the domain is restricted.

GCL verification state

Source lock: openai/math@1581d4cbdd69a0faf34b2089614fe2a814316a6a. The protected catalogue records one PDF, and the pinned publisher overview states the claimed superquadratic exponent and total-function scope.

This editorial pass has not checked the construction, calculated its global sensitivity and block sensitivity, or verified that the asserted exponent exceeds two after all parameter conversions. No Family 132-specific accepted critical proof audit, Lean trust-chain attestation, Comparator result, or MATHCERT certificate is asserted.

What remains

Read the pinned PDF and extract the exact family parameters, truth-table or algorithmic description, and complete theorem quantifiers. Reconstruct upper bounds for sensitivity and lower bounds for block sensitivity, checking that sensitive blocks are disjoint at a single input rather than merely across different inputs.

Determine whether floors, ceilings or sparse parameter sequences affect the exponent, and confirm all functions are total. Compare the result carefully with established sensitivity inequalities and the precise quadratic conjecture being targeted.

Complete mathematical Markdown/MkDocs regression, source-critical review and protected merge/readback. Editorial admission alone is not mathematical certification.

Sources and claim boundaries

Boundary: NO_CERTIFICATION. The superquadratic separation remains a publisher-reported claim pending independent verification.


Indexed primary manuscript paths: