{"abstract":"Start marker context is copied as its own target.","category":"Autoregressive target shifting","checks":6,"contract":"For observed tokens and bos, return [context,target] pairs beginning [bos, first token]; empty observations produce no pairs.","contract_signature":"tokens, bos","evaluation_group":"w3-ars-bos_prediction","failed_approach":"Dropping the marker entirely misaligns later positions when it should provide context.","family":"w3-autoregressive-target-shifting-bos-target-included","id":"FA-98251","implementations":{"attempt":{"sha256":"94f8a746603a65de983da0ea286f27b71a181261a71f8c3061a5f68ade7200ad","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(tokens, bos):\n    return [[tokens[i],tokens[i+1]] for i in range(len(tokens)-1)]\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\ncheck('first target',solve([N,N+1],0),[[0,N],[N,N+1]])\ncheck('one observed',solve([8],3),[[3,8]])\ncheck('empty',solve([],2),[])\ncheck('marker differs',solve([4,5],9),[[9,4],[4,5]])\ncheck('repeated token',solve([2,2],1),[[1,2],[2,2]])\ncheck('long stream',solve([1,2,3,4],7),[[7,1],[1,2],[2,3],[3,4]])\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":"eb6c56ede1f38be1d1dbcb23e994bbb1b04c82b4e653926774dfa7099d7d0792","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(tokens, bos):\n    s=[bos]+tokens\n    return [[s[i],s[i]] for i in range(len(tokens))]\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\ncheck('first target',solve([N,N+1],0),[[0,N],[N,N+1]])\ncheck('one observed',solve([8],3),[[3,8]])\ncheck('empty',solve([],2),[])\ncheck('marker differs',solve([4,5],9),[[9,4],[4,5]])\ncheck('repeated token',solve([2,2],1),[[1,2],[2,2]])\ncheck('long stream',solve([1,2,3,4],7),[[7,1],[1,2],[2,3],[3,4]])\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":"Inputs are toy integer token sequences; this contract says nothing about model quality, batching, sampling, or production training frameworks. 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-autoregressive-target-shifting-bos-target-included","generated_at":"2026-09-30T16:48:11.053892+00:00","license":"CC0-1.0","python":"3.12.14","seed":1,"split":"open-access"},"relevance":"A small deterministic token-stream contract makes next-token alignment errors inspectable.","root_cause":"The target at each position is copied from the same augmented-sequence index instead of the successor index.","sha256":"5e1aa9cc5084a8381c279e0b5dca2db5dfb8fa4e26b3b1a3958bbf28716bf4f6","title":"Start marker context is copied as its own target · 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":26.71,"exit_code":1,"observations":[{"actual":[[1,2]],"check":"first target","expected":[[0,1],[1,2]],"passed":false},{"actual":[],"check":"one observed","expected":[[3,8]],"passed":false},{"actual":[],"check":"empty","expected":[],"passed":true},{"actual":[[4,5]],"check":"marker differs","expected":[[9,4],[4,5]],"passed":false},{"actual":[[2,2]],"check":"repeated token","expected":[[1,2],[2,2]],"passed":false},{"actual":[[1,2],[2,3],[3,4]],"check":"long stream","expected":[[7,1],[1,2],[2,3],[3,4]],"passed":false}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"first target\", \"actual\": [[1, 2]], \"expected\": [[0, 1], [1, 2]], \"passed\": false}, {\"check\": \"one observed\", \"actual\": [], \"expected\": [[3, 8]], \"passed\": false}, {\"check\": \"empty\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"marker differs\", \"actual\": [[4, 5]], \"expected\": [[9, 4], [4, 5]], \"passed\": false}, {\"check\": \"repeated token\", \"actual\": [[2, 2]], \"expected\": [[1, 2], [2, 2]], \"passed\": false}, {\"check\": \"long stream\", \"actual\": [[1, 2], [2, 3], [3, 4]], \"expected\": [[7, 1], [1, 2], [2, 3], [3, 4]], \"passed\": false}], \"passed\": false}\n"},"broken":{"elapsed_ms":26.115,"exit_code":1,"observations":[{"actual":[[0,0],[1,1]],"check":"first target","expected":[[0,1],[1,2]],"passed":false},{"actual":[[3,3]],"check":"one observed","expected":[[3,8]],"passed":false},{"actual":[],"check":"empty","expected":[],"passed":true},{"actual":[[9,9],[4,4]],"check":"marker differs","expected":[[9,4],[4,5]],"passed":false},{"actual":[[1,1],[2,2]],"check":"repeated token","expected":[[1,2],[2,2]],"passed":false},{"actual":[[7,7],[1,1],[2,2],[3,3]],"check":"long stream","expected":[[7,1],[1,2],[2,3],[3,4]],"passed":false}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"first target\", \"actual\": [[0, 0], [1, 1]], \"expected\": [[0, 1], [1, 2]], \"passed\": false}, {\"check\": \"one observed\", \"actual\": [[3, 3]], \"expected\": [[3, 8]], \"passed\": false}, {\"check\": \"empty\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"marker differs\", \"actual\": [[9, 9], [4, 4]], \"expected\": [[9, 4], [4, 5]], \"passed\": false}, {\"check\": \"repeated token\", \"actual\": [[1, 1], [2, 2]], \"expected\": [[1, 2], [2, 2]], \"passed\": false}, {\"check\": \"long stream\", \"actual\": [[7, 7], [1, 1], [2, 2], [3, 3]], \"expected\": [[7, 1], [1, 2], [2, 3], [3, 4]], \"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."}}