{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:YGGXKUIEVTLSUF2R36O447M2MR","short_pith_number":"pith:YGGXKUIE","schema_version":"1.0","canonical_sha256":"c18d755104acd72a1751df9dce7d9a644468ebdeb1dde7e5d907cc0a0df4b46d","source":{"kind":"arxiv","id":"2608.12150","version":1},"attestation_state":"computed","paper":{"title":"Who Thinks Best Depends on How Long You Let Them: Budget-Dependent Rankings in LLM Evaluation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Alison R. Panisson, Rodrigo Guedes de Souza","submitted_at":"2026-08-12T15:11:35Z","abstract_excerpt":"Standard evaluation of large language models assumes stable model rankings across inference conditions. We challenge this assumption by varying the token generation budget, i.e., the maximum tokens a model may produce, across seven levels (64--4,096), evaluating four models on three reasoning benchmarks (56,476 inferences). We report four findings: (i) 3--19% of items exhibit non-monotone behavior (accuracy decreasing with more budget), even after controlling for truncation, and this phenomenon is model-specific (cross-model overlap: 6--14%). (ii) Model rankings reverse across budgets on all b"},"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":"2608.12150","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-08-12T15:11:35Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"08a13543f79fda7c7d25dd7ba19fe430de47acaadff6f19c7a623f92242f8cf6","abstract_canon_sha256":"b8820db3a9c5e4437505978bbb623330209e7f97e144dc221d184bd43043de22"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-13T01:30:14.347711Z","signature_b64":"S9oibrbuRYxauTfn/n1wO6giMitwY4xynLHt/joKRwMrYjq7/Dyh4fXeJkXPrYS3DyBPXLNKRj65h9u3tSo7DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c18d755104acd72a1751df9dce7d9a644468ebdeb1dde7e5d907cc0a0df4b46d","last_reissued_at":"2026-08-13T01:30:14.345013Z","signature_status":"signed_v1","first_computed_at":"2026-08-13T01:30:14.345013Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Who Thinks Best Depends on How Long You Let Them: Budget-Dependent Rankings in LLM Evaluation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Alison R. Panisson, Rodrigo Guedes de Souza","submitted_at":"2026-08-12T15:11:35Z","abstract_excerpt":"Standard evaluation of large language models assumes stable model rankings across inference conditions. We challenge this assumption by varying the token generation budget, i.e., the maximum tokens a model may produce, across seven levels (64--4,096), evaluating four models on three reasoning benchmarks (56,476 inferences). We report four findings: (i) 3--19% of items exhibit non-monotone behavior (accuracy decreasing with more budget), even after controlling for truncation, and this phenomenon is model-specific (cross-model overlap: 6--14%). (ii) Model rankings reverse across budgets on all b"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.12150","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/2608.12150/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":"2608.12150","created_at":"2026-08-13T01:30:14.346234+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.12150v1","created_at":"2026-08-13T01:30:14.346234+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.12150","created_at":"2026-08-13T01:30:14.346234+00:00"},{"alias_kind":"pith_short_12","alias_value":"YGGXKUIEVTLS","created_at":"2026-08-13T01:30:14.346234+00:00"},{"alias_kind":"pith_short_16","alias_value":"YGGXKUIEVTLSUF2R","created_at":"2026-08-13T01:30:14.346234+00:00"},{"alias_kind":"pith_short_8","alias_value":"YGGXKUIE","created_at":"2026-08-13T01:30:14.346234+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/YGGXKUIEVTLSUF2R36O447M2MR","json":"https://pith.science/pith/YGGXKUIEVTLSUF2R36O447M2MR.json","graph_json":"https://pith.science/api/pith-number/YGGXKUIEVTLSUF2R36O447M2MR/graph.json","events_json":"https://pith.science/api/pith-number/YGGXKUIEVTLSUF2R36O447M2MR/events.json","paper":"https://pith.science/paper/YGGXKUIE"},"agent_actions":{"view_html":"https://pith.science/pith/YGGXKUIEVTLSUF2R36O447M2MR","download_json":"https://pith.science/pith/YGGXKUIEVTLSUF2R36O447M2MR.json","view_paper":"https://pith.science/paper/YGGXKUIE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.12150&json=true","fetch_graph":"https://pith.science/api/pith-number/YGGXKUIEVTLSUF2R36O447M2MR/graph.json","fetch_events":"https://pith.science/api/pith-number/YGGXKUIEVTLSUF2R36O447M2MR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YGGXKUIEVTLSUF2R36O447M2MR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YGGXKUIEVTLSUF2R36O447M2MR/action/storage_attestation","attest_author":"https://pith.science/pith/YGGXKUIEVTLSUF2R36O447M2MR/action/author_attestation","sign_citation":"https://pith.science/pith/YGGXKUIEVTLSUF2R36O447M2MR/action/citation_signature","submit_replication":"https://pith.science/pith/YGGXKUIEVTLSUF2R36O447M2MR/action/replication_record"}},"created_at":"2026-08-13T01:30:14.346234+00:00","updated_at":"2026-08-13T01:30:14.346234+00:00"}