{"abstract":"Eligibility filtering runs after router top-k truncation.","category":"Mixture of experts dispatch","checks":6,"contract":"Routes [expert,score,eligible]; return up to k eligible routes by score descending then expert ID.","contract_signature":"routes,k","evaluation_group":"moe-filter-before-topk","failed_approach":"Filtering after top-k and refilling beyond the requested count admits one extra candidate.","family":"w3-mixture-of-experts-dispatch-postfilter-topk","id":"FA-100616","implementations":{"attempt":{"sha256":"c245e205d6e455f8374f16bd2aad0f59cd5a199281df9e927bae90381388d05c","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(routes,k):\n    return sorted([r for r in routes if r[2]],key=lambda x:(-x[1],x[0]))[:k+1]\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\nrows=[('a',N+3,False),('b',N+2,True),('c',N+1,True)]\ncheck('refill after ineligible',solve(rows,2),[('b',N+2,True),('c',N+1,True)])\ncheck('single eligible',solve(rows,1),[('b',N+2,True)])\ncheck('zero k',solve(rows,0),[])\ncheck('empty',solve([],N),[])\ncheck('none eligible',solve([('x',N,False)],2),[])\ncheck('all eligible ranked',solve([('z',N,True),('a',N+1,True)],2),[('a',N+1,True),('z',N,True)])\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":"83e1154a88559dbc152d3f40a558a1b13c57864e83b6dd2eb382e3888575b8b0","source":"\"\"\"Failure Map reference implementation. Python standard library only.\"\"\"\nimport json\n\nN = 1\nobservations = []\ndef solve(routes,k):\n    return [r for r in sorted(routes,key=lambda x:(-x[1],x[0]))[:k] if r[2]]\ndef check(label, actual, expected):\n    observations.append({\"check\": label, \"actual\": actual, \"expected\": expected, \"passed\": actual == expected})\nrows=[('a',N+3,False),('b',N+2,True),('c',N+1,True)]\ncheck('refill after ineligible',solve(rows,2),[('b',N+2,True),('c',N+1,True)])\ncheck('single eligible',solve(rows,1),[('b',N+2,True)])\ncheck('zero k',solve(rows,0),[])\ncheck('empty',solve([],N),[])\ncheck('none eligible',solve([('x',N,False)],2),[])\ncheck('all eligible ranked',solve([('z',N,True),('a',N+1,True)],2),[('a',N+1,True),('z',N,True)])\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 finite toy scores and routes; the contract does not model a production model, accelerator, or learned routing 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-mixture-of-experts-dispatch-postfilter-topk","generated_at":"2026-09-30T16:48:37.646430+00:00","license":"CC0-1.0","python":"3.12.14","seed":1,"split":"open-access"},"relevance":"A deterministic toy router makes this dispatch decision independently inspectable.","root_cause":"The candidate budget is consumed by unavailable experts before eligibility is applied.","sha256":"56b28e60c2420f0f2f99009d35fa40956b1cd98e29cacb2fd66284c2e11b826a","title":"Eligibility filtering runs after router top-k truncation · 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":30.792,"exit_code":1,"observations":[{"actual":[["b",3,true],["c",2,true]],"check":"refill after ineligible","expected":[["b",3,true],["c",2,true]],"passed":true},{"actual":[["b",3,true],["c",2,true]],"check":"single eligible","expected":[["b",3,true]],"passed":false},{"actual":[["b",3,true]],"check":"zero k","expected":[],"passed":false},{"actual":[],"check":"empty","expected":[],"passed":true},{"actual":[],"check":"none eligible","expected":[],"passed":true},{"actual":[["a",2,true],["z",1,true]],"check":"all eligible ranked","expected":[["a",2,true],["z",1,true]],"passed":true}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"refill after ineligible\", \"actual\": [[\"b\", 3, true], [\"c\", 2, true]], \"expected\": [[\"b\", 3, true], [\"c\", 2, true]], \"passed\": true}, {\"check\": \"single eligible\", \"actual\": [[\"b\", 3, true], [\"c\", 2, true]], \"expected\": [[\"b\", 3, true]], \"passed\": false}, {\"check\": \"zero k\", \"actual\": [[\"b\", 3, true]], \"expected\": [], \"passed\": false}, {\"check\": \"empty\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"none eligible\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"all eligible ranked\", \"actual\": [[\"a\", 2, true], [\"z\", 1, true]], \"expected\": [[\"a\", 2, true], [\"z\", 1, true]], \"passed\": true}], \"passed\": false}\n"},"broken":{"elapsed_ms":28.572,"exit_code":1,"observations":[{"actual":[["b",3,true]],"check":"refill after ineligible","expected":[["b",3,true],["c",2,true]],"passed":false},{"actual":[],"check":"single eligible","expected":[["b",3,true]],"passed":false},{"actual":[],"check":"zero k","expected":[],"passed":true},{"actual":[],"check":"empty","expected":[],"passed":true},{"actual":[],"check":"none eligible","expected":[],"passed":true},{"actual":[["a",2,true],["z",1,true]],"check":"all eligible ranked","expected":[["a",2,true],["z",1,true]],"passed":true}],"passed":false,"stderr":"","stdout":"{\"observations\": [{\"check\": \"refill after ineligible\", \"actual\": [[\"b\", 3, true]], \"expected\": [[\"b\", 3, true], [\"c\", 2, true]], \"passed\": false}, {\"check\": \"single eligible\", \"actual\": [], \"expected\": [[\"b\", 3, true]], \"passed\": false}, {\"check\": \"zero k\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"empty\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"none eligible\", \"actual\": [], \"expected\": [], \"passed\": true}, {\"check\": \"all eligible ranked\", \"actual\": [[\"a\", 2, true], [\"z\", 1, true]], \"expected\": [[\"a\", 2, true], [\"z\", 1, true]], \"passed\": true}], \"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."}}