verified certificate script asserted: CERTIFIED chi(qubit_sqrtT^2) >= 3, by an argument declared exact throughout (no floating-point margin)
1 run; 0.003 CPU-hours; 0.003 h wall-clock; on Apple silicon laptop, one core; about 10 s for the certificate; claude-fable-5-1, found the descent argument and wrote the certificate.
Settles χ(|sqrtT〉⊗2) = 3 with the three-term witness. Galois descent: a stabilizer span containing |sqrtT〉⊗2 contains its four conjugates (1, zeta^k)^2, k = 1, 5, 9, 13, four distinct points on the conic of product states, which span the three-dimensional symmetric subspace, so three terms are needed. Exact: the rank is computed in Q(zeta_16) and no margin stands under the bound. The numerical rank-2 exclusion over the 60 two-qubit states (closest approach 0.986 to the parallel threshold) agrees.
Vincent Russo · stabrank verify_challenge · Galois descent over Z[i]: the four conjugates of the target under Gal(Q(zeta_16)/Q(i)) span a space of dimension min(4, m + 1); no search
{
"schema_version": "0.1",
"orbit": "qubit_sqrtT",
"m": 2,
"direction": "lower",
"rank": 3,
"certificate": {
"script": "verify_challenge/cert_sqrtt_galois.py",
"expect": "CERTIFIED chi(qubit_sqrtT^2) >= 3",
"exact": true
},
"provenance": {
"author": "Vincent Russo",
"reference": "stabrank verify_challenge",
"method": "Galois descent over Z[i]: the four conjugates of the target under Gal(Q(zeta_16)/Q(i)) span a space of dimension min(4, m + 1); no search",
"date": "2026-10-21",
"github": [
"vprusso"
],
"compute": {
"cpu_hours": 0.003,
"wall_clock_hours": 0.003,
"runs": 1,
"hardware": "Apple silicon laptop, one core; about 10 s for the certificate",
"llm": [
{
"model": "claude-fable-5-1",
"role": "found the descent argument and wrote the certificate"
}
]
}
},
"notes": "Settles chi(|sqrtT>^2) = 3 with the three-term witness. Galois descent: a stabilizer span containing |sqrtT>^2 contains its four conjugates (1, zeta^k)^2, k = 1, 5, 9, 13, four distinct points on the conic of product states, which span the three-dimensional symmetric subspace, so three terms are needed. Exact: the rank is computed in Q(zeta_16) and no margin stands under the bound. The numerical rank-2 exclusion over the 60 two-qubit states (closest approach 0.986 to the parallel threshold) agrees."
}