FA-100076 / Preference pair construction / Open access

Preference pairs: pair weights remain unnormalized · case 01

pair weights remain unnormalized.

Verified by executionVariant 1 · 6 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

Raw annotator weights make the batch total depend on arbitrary weight scale.

THE FAILURE

Raw annotator weights make the batch total depend on arbitrary weight scale.

Unsuccessful approach: Normalizing by the number of rows ignores unequal annotator confidence.

Case contract

Input weights are nonnegative integers; return each positive weight divided by their positive sum as exact fractions [numerator,denominator], or [] if none are positive.

Why this case matters

A small deterministic preference-data transform makes pair semantics inspectable without a training framework.

1 / The failure

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(weights):
    return [[w,1] for w in weights if w>0]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('unequal weights',solve([N,2*N]),[[1,3],[2,3]])
check('zeros omitted',solve([0,N,0]),[[1,1]])
check('three-way',solve([1,1,2]),[[1,4],[1,4],[1,2]])
check('empty',solve([]),[])
check('all zero',solve([0,0]),[])
check('single positive',solve([N,0]),[[1,1]])
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

Boundary fixture	Actual	Expected	Outcome
unequal weights	`[[1, 1], [2, 1]]`	`[[1, 3], [2, 3]]`	Failed
zeros omitted	`[[1, 1]]`	`[[1, 1]]`	Passed
three-way	`[[1, 1], [1, 1], [2, 1]]`	`[[1, 4], [1, 4], [1, 2]]`	Failed
empty	`[]`	`[]`	Passed
all zero	`[]`	`[]`	Passed
single positive	`[[1, 1]]`	`[[1, 1]]`	Passed

SHA-256 / 220d3dcedf9401949fd78d8ca1e59a9e666e79122d9b434c33b699048cd7dbb3

2 / The unsuccessful fix

Exit 1

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(weights):
    return [[w,len([x for x in weights if x>0])] for w in weights if w>0]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('unequal weights',solve([N,2*N]),[[1,3],[2,3]])
check('zeros omitted',solve([0,N,0]),[[1,1]])
check('three-way',solve([1,1,2]),[[1,4],[1,4],[1,2]])
check('empty',solve([]),[])
check('all zero',solve([0,0]),[])
check('single positive',solve([N,0]),[[1,1]])
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

Boundary fixture	Actual	Expected	Outcome
unequal weights	`[[1, 2], [2, 2]]`	`[[1, 3], [2, 3]]`	Failed
zeros omitted	`[[1, 1]]`	`[[1, 1]]`	Passed
three-way	`[[1, 3], [1, 3], [2, 3]]`	`[[1, 4], [1, 4], [1, 2]]`	Failed
empty	`[]`	`[]`	Passed
all zero	`[]`	`[]`	Passed
single positive	`[[1, 1]]`	`[[1, 1]]`	Passed

SHA-256 / 60de7f2410a603ed8effdabaf1d1fb3f51d1f762886f1f5658a35493d30e970f

HELD IN THE MEMBER ARCHIVE

The verified repair and its recorded checks are member-only.

This mechanism has 6 recorded checks per implementation. The open-access tier publishes the failure and the unsuccessful fix; the repaired source that passes every check, and the observations that prove it, are available to members.

Every case sharing this mechanism uses the same contract and the same repair, so this one record is held back for all of them.

Member access is invitation-based. Sign in with your invited account to inspect the repair.

Verification & scope

These are explicit toy record contracts; they do not claim compatibility with a trainer or improved model behavior. This reproducer isolates one failure mechanism. Results cover the supplied fixtures. Variants within a family share a test contract and should remain grouped when constructing evaluation splits. Related mechanisms with a shared evaluation_group must also remain together; these controlled models are not independent production incidents.

Observations recorded using Python 3.12.14 at 2026-09-30T16:48:31.292970+00:00.

Case digest / eeff035ebad08e533f74995a0f362ac7d75d404a09abadd4d195c6da8c1dc4f8