FA-98151 / Attention mask construction / Open access

Padded query rows can attend to real keys · case 01

Padded query rows can attend to real keys.

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

ROOT CAUSE

Only key padding is considered, so invalid query rows are populated.

THE FAILURE

Only key padding is considered, so invalid query rows are populated.

Unsuccessful approach: Masking only query validity still exposes padded key positions from a valid query.

Case contract

Square matrix; pair visibility requires query valid, key valid and key<=query.

Why this case matters

A small explicit visibility matrix makes sequence-attention boundary behavior reviewable.

1 / The failure

Exit 1

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

N = 1
observations = []
def solve(valid):
    return [[bool(valid[j]) and j<=i for j in range(len(valid))] for i in range(len(valid))]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
v=[True]*N+[False]
check('padded query row',solve(v),[[bool(v[i]) and bool(v[j]) and j<=i for j in range(N+1)] for i in range(N+1)])
check('active query',solve([True,True]),[[True,False],[True,True]])
check('all invalid',solve([False,False]),[[False,False],[False,False]])
check('empty',solve([]),[])
check('singleton invalid',solve([False]),[[False]])
check('interior pad',solve([True,False,True]),[[True,False,False],[False,False,False],[True,False,True]])
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
padded query row	`[[True, False], [True, False]]`	`[[True, False], [False, False]]`	Failed
active query	`[[True, False], [True, True]]`	`[[True, False], [True, True]]`	Passed
all invalid	`[[False, False], [False, False]]`	`[[False, False], [False, False]]`	Passed
empty	`[]`	`[]`	Passed
singleton invalid	`[[False]]`	`[[False]]`	Passed
interior pad	`[[True, False, False], [True, False, False], [True, False, True]]`	`[[True, False, False], [False, False, False], [True, False, True]]`	Failed

SHA-256 / 29b4af6e1b85fdedf701c5e5b7988accdc80fe1bb039c79420c9c27607886fd7

2 / The unsuccessful fix

Exit 1

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

N = 1
observations = []
def solve(valid):
    return [[bool(valid[i]) and j<=i for j in range(len(valid))] for i in range(len(valid))]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
v=[True]*N+[False]
check('padded query row',solve(v),[[bool(v[i]) and bool(v[j]) and j<=i for j in range(N+1)] for i in range(N+1)])
check('active query',solve([True,True]),[[True,False],[True,True]])
check('all invalid',solve([False,False]),[[False,False],[False,False]])
check('empty',solve([]),[])
check('singleton invalid',solve([False]),[[False]])
check('interior pad',solve([True,False,True]),[[True,False,False],[False,False,False],[True,False,True]])
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
padded query row	`[[True, False], [False, False]]`	`[[True, False], [False, False]]`	Passed
active query	`[[True, False], [True, True]]`	`[[True, False], [True, True]]`	Passed
all invalid	`[[False, False], [False, False]]`	`[[False, False], [False, False]]`	Passed
empty	`[]`	`[]`	Passed
singleton invalid	`[[False]]`	`[[False]]`	Passed
interior pad	`[[True, False, False], [False, False, False], [True, True, True]]`	`[[True, False, False], [False, False, False], [True, False, True]]`	Failed

SHA-256 / fd34e97425eca81a1a0eebc6f8544ecaaf01c90dfc8d0f90fa3142b729ef0aeb

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 toy Boolean matrix contracts; they do not define a production model architecture or tensor library 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 2026-09-30T16:48:09.757508+00:00.

Case digest / 19f6450e84f76799c707434ed2071f3c9fe55e78e622c012c018e2cc70ba69d1