FA-98066 / Sequence batch packing / Open access

Separator tokens are omitted from the packing budget · case 01

Separator tokens are omitted from the packing budget.

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

ROOT CAUSE

The accounting omits the inter-sequence separator cost.

THE FAILURE

The accounting omits the inter-sequence separator cost.

Unsuccessful approach: Charging two separators splits a bin that fits with exactly one.

Case contract

Pack in order, counting one separator token between adjacent sequences in each bin; output bins preserve original sequences.

Why this case matters

This deterministic toy packing rule models a common boundary in preparing bounded training sequences.

1 / The failure

Exit 1

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

N = 1
observations = []
def solve(sequences, capacity, separator):
    bins=[]; cur=[]; used=0
    for s in sequences:
        cost=len(s)
        if cur and used+cost>capacity: bins.append(cur); cur=[]; used=0; cost=len(s)
        cur.append(s); used+=cost
    if cur: bins.append(cur)
    return bins
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
a=[1]*N
check('separator tips over', solve([a,[2]],N+1,99), [[a],[[2]]])
check('separator exact', solve([a,[2]],N+2,99), [[a,[2]]])
check('single no separator', solve([a],N,99), [[a]])
check('empty', solve([],N,99), [])
check('three sequences', solve([[1],[2],[3]],5,99), [[[1],[2],[3]]])
check('independent bins', solve([[1]*N,[2]*N],N,99), [[[1]*N],[[2]*N]])
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
separator tips over	`[[[1], [2]]]`	`[[[1]], [[2]]]`	Failed
separator exact	`[[[1], [2]]]`	`[[[1], [2]]]`	Passed
single no separator	`[[[1]]]`	`[[[1]]]`	Passed
empty	`[]`	`[]`	Passed
three sequences	`[[[1], [2], [3]]]`	`[[[1], [2], [3]]]`	Passed
independent bins	`[[[1]], [[2]]]`	`[[[1]], [[2]]]`	Passed

SHA-256 / 074ce8b8964dc69642b4e10787628beaae8c242bcfbdd6da9ee73034fcdaa733

2 / The unsuccessful fix

Exit 1

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

N = 1
observations = []
def solve(sequences, capacity, separator):
    bins=[]; cur=[]; used=0
    for s in sequences:
        cost=len(s)+(2 if cur else 0)
        if cur and used+cost>capacity: bins.append(cur); cur=[]; used=0; cost=len(s)
        cur.append(s); used+=cost
    if cur: bins.append(cur)
    return bins
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
a=[1]*N
check('separator tips over', solve([a,[2]],N+1,99), [[a],[[2]]])
check('separator exact', solve([a,[2]],N+2,99), [[a,[2]]])
check('single no separator', solve([a],N,99), [[a]])
check('empty', solve([],N,99), [])
check('three sequences', solve([[1],[2],[3]],5,99), [[[1],[2],[3]]])
check('independent bins', solve([[1]*N,[2]*N],N,99), [[[1]*N],[[2]*N]])
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
separator tips over	`[[[1]], [[2]]]`	`[[[1]], [[2]]]`	Passed
separator exact	`[[[1]], [[2]]]`	`[[[1], [2]]]`	Failed
single no separator	`[[[1]]]`	`[[[1]]]`	Passed
empty	`[]`	`[]`	Passed
three sequences	`[[[1], [2]], [[3]]]`	`[[[1], [2], [3]]]`	Failed
independent bins	`[[[1]], [[2]]]`	`[[[1]], [[2]]]`	Passed

SHA-256 / 2922f332e536906b15a35b83330d6c4afc29739ad657937b3642c0bd74ccccdc

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

Inputs are short integer-token lists under the stated toy contract; results do not claim compatibility with a specific training framework. 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 2026-09-30T16:48:08.896751+00:00.

Case digest / 96ef5f7268bdabfb2e4c58f209c2c415b432f63163bccc8c7451dd5159ab9f3e