{"abstract":"Committed score includes the rejected token score.","category":"Speculative decoding bookkeeping","checks":7,"contract":"Sum scores at positions before accepted count k.","contract_signature":"scores, accepted","evaluation_group":"spec-accepted-score-sum","failed_approach":"Using only the first score ignores later accepted positions.","family":"w3-speculative-decoding-bookkeeping-accepted-score-sum","id":"FA-100731","implementations":{"attempt":{"sha256":"16b57e3eb522b748b24be1b453ab91c323bfd0704a5c390e4586994ed50f0822","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(scores, accepted):\n    return scores[0] if accepted else 0\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\ncheck('partial scores', solve([2,3,7],2),5)\ncheck('none accepted', solve([N],0),0)\ncheck('all accepted', solve([1,4],2),5)\ncheck('empty scores', solve([],0),0)\ncheck('single accepted', solve([N],1),N)\ncheck('zero-valued token', solve([0,5],1),0)\ncheck('skip rejected tail', solve([1,2,9],2),3)\nprint(json.dumps({\"observations\": observations, \"passed\": all(x[\"passed\"] for x in observations)}, ensure_ascii=False))\nraise SystemExit(0 if all(x[\"passed\"] for x in observations) else 1)\n"},"broken":{"sha256":"5c8fdaf4892f6cda2c0d69206a7ee6c529035199d90d58660591907141ea0e50","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(scores, accepted):\n    return sum(scores[:accepted+1])\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\ncheck('partial scores', solve([2,3,7],2),5)\ncheck('none accepted', solve([N],0),0)\ncheck('all accepted', solve([1,4],2),5)\ncheck('empty scores', solve([],0),0)\ncheck('single accepted', solve([N],1),N)\ncheck('zero-valued token', solve([0,5],1),0)\ncheck('skip rejected tail', solve([1,2,9],2),3)\nprint(json.dumps({\"observations\": observations, \"passed\": all(x[\"passed\"] for x in observations)}, ensure_ascii=False))\nraise SystemExit(0 if all(x[\"passed\"] for x in observations) else 1)\n"}},"limitations":"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.","method":"Deterministic executable model with adversarial boundary fixtures.","provenance":{"created_by":"Failure Map","dependencies":"Python standard library","family":"w3-speculative-decoding-bookkeeping-accepted-score-sum","generated_at":"2026-09-30T16:48:39.096437+00:00","license":"CC0-1.0","python":"3.12.14","seed":1,"split":"open-access"},"relevance":"A self-contained bookkeeping model makes prefix and token-count semantics observable without a model runtime.","root_cause":"The rejected token score is included in committed score.","sha256":"c2347a458e8e257b5d1a06101d19c53b89340fd1bcc73e2a25666cf0c315583c","title":"Decoding: Committed score includes the rejected token score · case 01","variant":1,"variant_policy":"The fixture code references N, supplied as 1 through 5. Fixed boundary checks may repeat. These records share one mechanism and must remain together in evaluation splits.","verified":true,"visibility":"public","verification":{"attempt":{"elapsed_ms":33.786,"exit_code":1,"observations":[{"actual":2,"check":"partial scores","expected":5,"passed":false},{"actual":0,"check":"none accepted","expected":0,"passed":true},{"actual":1,"check":"all accepted","expected":5,"passed":false},{"actual":0,"check":"empty scores","expected":0,"passed":true},{"actual":1,"check":"single accepted","expected":1,"passed":true},{"actual":0,"check":"zero-valued token","expected":0,"passed":true},{"actual":1,"check":"skip rejected tail","expected":3,"passed":false}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"partial scores\", \"actual\": 2, \"expected\": 5, \"passed\": false}, {\"check\": \"none accepted\", \"actual\": 0, \"expected\": 0, \"passed\": true}, {\"check\": \"all accepted\", \"actual\": 1, \"expected\": 5, \"passed\": false}, {\"check\": \"empty scores\", \"actual\": 0, \"expected\": 0, \"passed\": true}, {\"check\": \"single accepted\", \"actual\": 1, \"expected\": 1, \"passed\": true}, {\"check\": \"zero-valued token\", \"actual\": 0, \"expected\": 0, \"passed\": true}, {\"check\": \"skip rejected tail\", \"actual\": 1, \"expected\": 3, \"passed\": false}], \"passed\": false}\n"},"broken":{"elapsed_ms":34.433,"exit_code":1,"observations":[{"actual":12,"check":"partial scores","expected":5,"passed":false},{"actual":1,"check":"none accepted","expected":0,"passed":false},{"actual":5,"check":"all accepted","expected":5,"passed":true},{"actual":0,"check":"empty scores","expected":0,"passed":true},{"actual":1,"check":"single accepted","expected":1,"passed":true},{"actual":5,"check":"zero-valued token","expected":0,"passed":false},{"actual":12,"check":"skip rejected tail","expected":3,"passed":false}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"partial scores\", \"actual\": 12, \"expected\": 5, \"passed\": false}, {\"check\": \"none accepted\", \"actual\": 1, \"expected\": 0, \"passed\": false}, {\"check\": \"all accepted\", \"actual\": 5, \"expected\": 5, \"passed\": true}, {\"check\": \"empty scores\", \"actual\": 0, \"expected\": 0, \"passed\": true}, {\"check\": \"single accepted\", \"actual\": 1, \"expected\": 1, \"passed\": true}, {\"check\": \"zero-valued token\", \"actual\": 5, \"expected\": 0, \"passed\": false}, {\"check\": \"skip rejected tail\", \"actual\": 12, \"expected\": 3, \"passed\": false}], \"passed\": false}\n"}},"member_only":{"stages":["fixed"],"fields":["implementations.fixed","verification.fixed","harness","repair"],"note":"The verified repair, its recorded checks, the repair description, and the scoring harness are available to members."}}