{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:TQXWJUBRKUW44XDQPHAHBSSZAO","short_pith_number":"pith:TQXWJUBR","schema_version":"1.0","canonical_sha256":"9c2f64d031552dce5c7079c070ca590387be268a41c250ecdc34beab1a9e04a8","source":{"kind":"arxiv","id":"2509.11208","version":3},"attestation_state":"computed","paper":{"title":"Predictable Compression Failures: Order Sensitivity and Information Budgeting for Evidence-Grounded Binary Adjudication","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Ahmed Karim, Leon Chlon, Maggie Chlon, MarcAntonio Awada","submitted_at":"2025-09-14T10:32:59Z","abstract_excerpt":"Transformers used for evidence-grounded binary adjudication (e.g., support/refute, yes/no, or verifier-backed pass/fail decisions) can be sensitive to the order in which exchangeable evidence is presented, producing dispersion across permutations and unreliable attempted answers under a verifier-relative Bernoulli predicate. We treat evidence order as a nuisance variable and formalize an expectation-realization gap: next-token training can minimize expected conditional description length over orderings while a fixed ordering remains position-sensitive. Our Quantified Martingale Violation (QMV)"},"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":"2509.11208","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2025-09-14T10:32:59Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"499b9ceebe5915ca07cef8113eb81058aa49e7ca81030bf64d70b7e6fda095f5","abstract_canon_sha256":"3ce9a8d109ef8f55f8e054529266a5b1e63123296e85c0d1287516e6d8aeb254"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-08T01:03:48.390271Z","signature_b64":"sUzNkH50B8Z2/O4S2kgrrCUwfd6Ly9pTl9i6A2//qEp0SfIVsrBZ57S/7/2AAvM5SYHOUuRvJjudi0FusFiIAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9c2f64d031552dce5c7079c070ca590387be268a41c250ecdc34beab1a9e04a8","last_reissued_at":"2026-06-08T01:03:48.389249Z","signature_status":"signed_v1","first_computed_at":"2026-06-08T01:03:48.389249Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Predictable Compression Failures: Order Sensitivity and Information Budgeting for Evidence-Grounded Binary Adjudication","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Ahmed Karim, Leon Chlon, Maggie Chlon, MarcAntonio Awada","submitted_at":"2025-09-14T10:32:59Z","abstract_excerpt":"Transformers used for evidence-grounded binary adjudication (e.g., support/refute, yes/no, or verifier-backed pass/fail decisions) can be sensitive to the order in which exchangeable evidence is presented, producing dispersion across permutations and unreliable attempted answers under a verifier-relative Bernoulli predicate. We treat evidence order as a nuisance variable and formalize an expectation-realization gap: next-token training can minimize expected conditional description length over orderings while a fixed ordering remains position-sensitive. Our Quantified Martingale Violation (QMV)"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.11208","kind":"arxiv","version":3},"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/2509.11208/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":"2509.11208","created_at":"2026-06-08T01:03:48.389399+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.11208v3","created_at":"2026-06-08T01:03:48.389399+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.11208","created_at":"2026-06-08T01:03:48.389399+00:00"},{"alias_kind":"pith_short_12","alias_value":"TQXWJUBRKUW4","created_at":"2026-06-08T01:03:48.389399+00:00"},{"alias_kind":"pith_short_16","alias_value":"TQXWJUBRKUW44XDQ","created_at":"2026-06-08T01:03:48.389399+00:00"},{"alias_kind":"pith_short_8","alias_value":"TQXWJUBR","created_at":"2026-06-08T01:03:48.389399+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2603.11749","citing_title":"Truth as a Compression Artifact in Language Model Training","ref_index":2,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TQXWJUBRKUW44XDQPHAHBSSZAO","json":"https://pith.science/pith/TQXWJUBRKUW44XDQPHAHBSSZAO.json","graph_json":"https://pith.science/api/pith-number/TQXWJUBRKUW44XDQPHAHBSSZAO/graph.json","events_json":"https://pith.science/api/pith-number/TQXWJUBRKUW44XDQPHAHBSSZAO/events.json","paper":"https://pith.science/paper/TQXWJUBR"},"agent_actions":{"view_html":"https://pith.science/pith/TQXWJUBRKUW44XDQPHAHBSSZAO","download_json":"https://pith.science/pith/TQXWJUBRKUW44XDQPHAHBSSZAO.json","view_paper":"https://pith.science/paper/TQXWJUBR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.11208&json=true","fetch_graph":"https://pith.science/api/pith-number/TQXWJUBRKUW44XDQPHAHBSSZAO/graph.json","fetch_events":"https://pith.science/api/pith-number/TQXWJUBRKUW44XDQPHAHBSSZAO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TQXWJUBRKUW44XDQPHAHBSSZAO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TQXWJUBRKUW44XDQPHAHBSSZAO/action/storage_attestation","attest_author":"https://pith.science/pith/TQXWJUBRKUW44XDQPHAHBSSZAO/action/author_attestation","sign_citation":"https://pith.science/pith/TQXWJUBRKUW44XDQPHAHBSSZAO/action/citation_signature","submit_replication":"https://pith.science/pith/TQXWJUBRKUW44XDQPHAHBSSZAO/action/replication_record"}},"created_at":"2026-06-08T01:03:48.389399+00:00","updated_at":"2026-06-08T01:03:48.389399+00:00"}