FA-98011 / Training label alignment / Open access

Student targets use teacher class indices directly · case 01

Student targets use teacher class indices directly.

Verified by executionVariant 1 · 8 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

The teacher and student class vocabularies use different index assignments.

THE FAILURE

The teacher and student class vocabularies use different index assignments.

Unsuccessful approach: Clamping teacher indices to the student range converts invalid classes into plausible labels.

Case contract

Map each teacher index through teacher_classes to the matching student_classes index; missing names or invalid indices yield -1.

Why this case matters

A small deterministic training-data operation makes label ownership and ordering explicit.

1 / The failure

Exit 1

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

N = 1
observations = []
def solve(teacher_indices, teacher_classes, student_classes):
    return [i if 0<=i<len(student_classes) else -1 for i in teacher_indices]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('different order',solve([0,1],['cat','dog'],['dog','cat']),[1,0])
check('missing student class',solve([0],['bird'],['cat']),[-1])
check('invalid teacher index',solve([3,-1],['a','b'],['a']),[-1,-1])
check('empty batch',solve([],['a'],['a']),[])
check('same singleton',solve([0],['x'],['x']),[0])
check('mixed mapping',solve([2,0,1],['a','b','c'],['c','a']),[0,1,-1])
check('empty student',solve([0],['a'],[]),[-1])
check('variant vocabulary order',solve([0,1],['label'+str(N),'base'],['base','label'+str(N)]),[1,0])
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
different order	`[0, 1]`	`[1, 0]`	Failed
missing student class	`[0]`	`[-1]`	Failed
invalid teacher index	`[-1, -1]`	`[-1, -1]`	Passed
empty batch	`[]`	`[]`	Passed
same singleton	`[0]`	`[0]`	Passed
mixed mapping	`[-1, 0, 1]`	`[0, 1, -1]`	Failed
empty student	`[-1]`	`[-1]`	Passed
variant vocabulary order	`[0, 1]`	`[1, 0]`	Failed

SHA-256 / ec8738ef760146e5ec6d2bbcc8c2d23b91bff3404c6413405bdc7b71c513951f

2 / The unsuccessful fix

Exit 1

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

N = 1
observations = []
def solve(teacher_indices, teacher_classes, student_classes):
    return [min(i,len(student_classes)-1) if student_classes and i>=0 else -1 for i in teacher_indices]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('different order',solve([0,1],['cat','dog'],['dog','cat']),[1,0])
check('missing student class',solve([0],['bird'],['cat']),[-1])
check('invalid teacher index',solve([3,-1],['a','b'],['a']),[-1,-1])
check('empty batch',solve([],['a'],['a']),[])
check('same singleton',solve([0],['x'],['x']),[0])
check('mixed mapping',solve([2,0,1],['a','b','c'],['c','a']),[0,1,-1])
check('empty student',solve([0],['a'],[]),[-1])
check('variant vocabulary order',solve([0,1],['label'+str(N),'base'],['base','label'+str(N)]),[1,0])
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
different order	`[0, 1]`	`[1, 0]`	Failed
missing student class	`[0]`	`[-1]`	Failed
invalid teacher index	`[0, -1]`	`[-1, -1]`	Failed
empty batch	`[]`	`[]`	Passed
same singleton	`[0]`	`[0]`	Passed
mixed mapping	`[1, 0, 1]`	`[0, 1, -1]`	Failed
empty student	`[-1]`	`[-1]`	Passed
variant vocabulary order	`[0, 1]`	`[1, 0]`	Failed

SHA-256 / 631cb4c73847abf6be8e31ea5e4beac1bb56be6d96789aa02a3bc4832222e0a2

HELD IN THE MEMBER ARCHIVE

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

This mechanism has 8 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

Class vocabularies contain unique names; indices are integers. 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 2026-09-30T16:48:08.407453+00:00.

Case digest / 95f8e24823b447478b8cb95cf02f879735c1f7a550b56f09146c4164dd42c942