{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:QNGUHUXXFFVZRZO7AQJCB6KFBH","short_pith_number":"pith:QNGUHUXX","schema_version":"1.0","canonical_sha256":"834d43d2f7296b98e5df041220f94509cbf77538d9a0efa504dde444e42a9762","source":{"kind":"arxiv","id":"2607.24268","version":1},"attestation_state":"computed","paper":{"title":"Accuracy Hides How Language Models Fail: Measuring Failure States Under Matched Output Budgets","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Yinghan Hou, Zongyou Yang","submitted_at":"2026-07-27T11:04:17Z","abstract_excerpt":"Language-model benchmarks collapse two distinct measurement questions into a single accuracy score: whether a response reached an evaluable state, and whether its answer was judged correct. We introduce a two-layer evaluation framework that separates scorer-independent execution evidence, including termination, answer exposure, parseability, and completion length, from scorer-dependent correctness. Across 2,550 outputs from five fixed Qwen and DeepSeek configurations on MATH and ARC-Challenge, matched 2,048-token limits produce sharply different execution mixtures: 49 of 450 Qwen MATH outputs "},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2607.24268","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-07-27T11:04:17Z","cross_cats_sorted":[],"title_canon_sha256":"a2a1d02c09676596a1d3a9c2b015e15a9f7b19c3aa8bf06d58b643d6112e939e","abstract_canon_sha256":"11298b943a7ed982a4cf01d6226055ffc3b7a4ce7ad95f647f11f19b9d75fa00"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-28T02:23:54.475923Z","signature_b64":"cz0vmkMJTaNMXdG+kTEg5g/HjLNlcmKY7V4/ZQipnKffg+IRTI6bEcKQWMZ9l3AHrkHuLwHid2vF6qee6CjJCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"834d43d2f7296b98e5df041220f94509cbf77538d9a0efa504dde444e42a9762","last_reissued_at":"2026-07-28T02:23:54.475115Z","signature_status":"signed_v1","first_computed_at":"2026-07-28T02:23:54.475115Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Accuracy Hides How Language Models Fail: Measuring Failure States Under Matched Output Budgets","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Yinghan Hou, Zongyou Yang","submitted_at":"2026-07-27T11:04:17Z","abstract_excerpt":"Language-model benchmarks collapse two distinct measurement questions into a single accuracy score: whether a response reached an evaluable state, and whether its answer was judged correct. We introduce a two-layer evaluation framework that separates scorer-independent execution evidence, including termination, answer exposure, parseability, and completion length, from scorer-dependent correctness. Across 2,550 outputs from five fixed Qwen and DeepSeek configurations on MATH and ARC-Challenge, matched 2,048-token limits produce sharply different execution mixtures: 49 of 450 Qwen MATH outputs "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.24268","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2607.24268/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2607.24268","created_at":"2026-07-28T02:23:54.475535+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.24268v1","created_at":"2026-07-28T02:23:54.475535+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.24268","created_at":"2026-07-28T02:23:54.475535+00:00"},{"alias_kind":"pith_short_12","alias_value":"QNGUHUXXFFVZ","created_at":"2026-07-28T02:23:54.475535+00:00"},{"alias_kind":"pith_short_16","alias_value":"QNGUHUXXFFVZRZO7","created_at":"2026-07-28T02:23:54.475535+00:00"},{"alias_kind":"pith_short_8","alias_value":"QNGUHUXX","created_at":"2026-07-28T02:23:54.475535+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QNGUHUXXFFVZRZO7AQJCB6KFBH","json":"https://pith.science/pith/QNGUHUXXFFVZRZO7AQJCB6KFBH.json","graph_json":"https://pith.science/api/pith-number/QNGUHUXXFFVZRZO7AQJCB6KFBH/graph.json","events_json":"https://pith.science/api/pith-number/QNGUHUXXFFVZRZO7AQJCB6KFBH/events.json","paper":"https://pith.science/paper/QNGUHUXX"},"agent_actions":{"view_html":"https://pith.science/pith/QNGUHUXXFFVZRZO7AQJCB6KFBH","download_json":"https://pith.science/pith/QNGUHUXXFFVZRZO7AQJCB6KFBH.json","view_paper":"https://pith.science/paper/QNGUHUXX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.24268&json=true","fetch_graph":"https://pith.science/api/pith-number/QNGUHUXXFFVZRZO7AQJCB6KFBH/graph.json","fetch_events":"https://pith.science/api/pith-number/QNGUHUXXFFVZRZO7AQJCB6KFBH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QNGUHUXXFFVZRZO7AQJCB6KFBH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QNGUHUXXFFVZRZO7AQJCB6KFBH/action/storage_attestation","attest_author":"https://pith.science/pith/QNGUHUXXFFVZRZO7AQJCB6KFBH/action/author_attestation","sign_citation":"https://pith.science/pith/QNGUHUXXFFVZRZO7AQJCB6KFBH/action/citation_signature","submit_replication":"https://pith.science/pith/QNGUHUXXFFVZRZO7AQJCB6KFBH/action/replication_record"}},"created_at":"2026-07-28T02:23:54.475535+00:00","updated_at":"2026-07-28T02:23:54.475535+00:00"}