FAILURE MAP
← Case archive

FA-100731 / Speculative decoding bookkeeping / Open access

Decoding: Committed score includes the rejected token score · case 01

Committed score includes the rejected token score.

Verified by executionVariant 1 · 7 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

The rejected token score is included in committed score.

THE FAILURE

The rejected token score is included in committed score.

Unsuccessful approach: Using only the first score ignores later accepted positions.

Case contract

Sum scores at positions before accepted count k.

Why this case matters

A self-contained bookkeeping model makes prefix and token-count semantics observable without a model runtime.

1 / The failure

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(scores, accepted):
    return sum(scores[:accepted+1])
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('partial scores', solve([2,3,7],2),5)
check('none accepted', solve([N],0),0)
check('all accepted', solve([1,4],2),5)
check('empty scores', solve([],0),0)
check('single accepted', solve([N],1),N)
check('zero-valued token', solve([0,5],1),0)
check('skip rejected tail', solve([1,2,9],2),3)
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 fixtureActualExpectedOutcome
partial scores125Failed
none accepted10Failed
all accepted55Passed
empty scores00Passed
single accepted11Passed
zero-valued token50Failed
skip rejected tail123Failed

SHA-256 / 5c8fdaf4892f6cda2c0d69206a7ee6c529035199d90d58660591907141ea0e50

2 / The unsuccessful fix

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(scores, accepted):
    return scores[0] if accepted else 0
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('partial scores', solve([2,3,7],2),5)
check('none accepted', solve([N],0),0)
check('all accepted', solve([1,4],2),5)
check('empty scores', solve([],0),0)
check('single accepted', solve([N],1),N)
check('zero-valued token', solve([0,5],1),0)
check('skip rejected tail', solve([1,2,9],2),3)
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 fixtureActualExpectedOutcome
partial scores25Failed
none accepted00Passed
all accepted15Failed
empty scores00Passed
single accepted11Passed
zero-valued token00Passed
skip rejected tail13Failed

SHA-256 / 16b57e3eb522b748b24be1b453ab91c323bfd0704a5c390e4586994ed50f0822

HELD IN THE MEMBER ARCHIVE

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

This mechanism has 7 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.

Sign in to the archive ↗

Verification & scope

These are finite toy sequences with explicit contracts; they make no claim about production decoder behavior or model quality. 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 2026-09-30T16:48:39.096437+00:00.

Case digest / c2347a458e8e257b5d1a06101d19c53b89340fd1bcc73e2a25666cf0c315583c