{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2W5SNTKHBIARME5HHRFFN7ATYP","short_pith_number":"pith:2W5SNTKH","schema_version":"1.0","canonical_sha256":"d5bb26cd470a011613a73c4a56fc13c3f5a9bc494297e969630145613293d096","source":{"kind":"arxiv","id":"2606.11196","version":1},"attestation_state":"computed","paper":{"title":"PoQ-Judge: A Multi-Architecture Evaluation Framework for Cost-Aware Proof-of-Quality in Decentralized LLM Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aaron Chan, Alex Ding, Arther Tian, Frank Chen, Simon Wu","submitted_at":"2026-04-20T02:29:52Z","abstract_excerpt":"Decentralized LLM inference networks need lightweight, reference-free quality evaluation for Proof of Quality (PoQ). We present PoQ-Judge, a framework that trains dedicated judge models to score query-output pairs without ground-truth references. We study three architectures across the quality-cost tradeoff: a TextCNN judge, a MiniLM cross-encoder, and a DeBERTa judge. Using two-stage training on UltraFeedback plus GPT-labeled in-domain data, the best model reaches 0.747 Pearson correlation with the ground-truth proxy on a held-out test set, outperforming reference-based evaluators from prior "},"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":"2606.11196","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-04-20T02:29:52Z","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"title_canon_sha256":"bbc259aa91ffb74a70372e147940e15666fd848658c632cbb7356cc39d224bda","abstract_canon_sha256":"3a073112b22babcbc40d5dd88606c60f776d3fbd17b2f6fd9e9bd810cf1527d5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-11T00:08:05.839614Z","signature_b64":"z+hC8cgB0euqV/LHE1b7bxzYZTH733r214Xrc60TeN+WexWsVNbLAOYBjcn8RYFQ0ob4HVyCexWe+ZTWhAVBAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d5bb26cd470a011613a73c4a56fc13c3f5a9bc494297e969630145613293d096","last_reissued_at":"2026-06-11T00:08:05.838631Z","signature_status":"signed_v1","first_computed_at":"2026-06-11T00:08:05.838631Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PoQ-Judge: A Multi-Architecture Evaluation Framework for Cost-Aware Proof-of-Quality in Decentralized LLM Inference","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Aaron Chan, Alex Ding, Arther Tian, Frank Chen, Simon Wu","submitted_at":"2026-04-20T02:29:52Z","abstract_excerpt":"Decentralized LLM inference networks need lightweight, reference-free quality evaluation for Proof of Quality (PoQ). We present PoQ-Judge, a framework that trains dedicated judge models to score query-output pairs without ground-truth references. We study three architectures across the quality-cost tradeoff: a TextCNN judge, a MiniLM cross-encoder, and a DeBERTa judge. Using two-stage training on UltraFeedback plus GPT-labeled in-domain data, the best model reaches 0.747 Pearson correlation with the ground-truth proxy on a held-out test set, outperforming reference-based evaluators from prior "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.11196","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/2606.11196/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":"2606.11196","created_at":"2026-06-11T00:08:05.838767+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.11196v1","created_at":"2026-06-11T00:08:05.838767+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.11196","created_at":"2026-06-11T00:08:05.838767+00:00"},{"alias_kind":"pith_short_12","alias_value":"2W5SNTKHBIAR","created_at":"2026-06-11T00:08:05.838767+00:00"},{"alias_kind":"pith_short_16","alias_value":"2W5SNTKHBIARME5H","created_at":"2026-06-11T00:08:05.838767+00:00"},{"alias_kind":"pith_short_8","alias_value":"2W5SNTKH","created_at":"2026-06-11T00:08:05.838767+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/2W5SNTKHBIARME5HHRFFN7ATYP","json":"https://pith.science/pith/2W5SNTKHBIARME5HHRFFN7ATYP.json","graph_json":"https://pith.science/api/pith-number/2W5SNTKHBIARME5HHRFFN7ATYP/graph.json","events_json":"https://pith.science/api/pith-number/2W5SNTKHBIARME5HHRFFN7ATYP/events.json","paper":"https://pith.science/paper/2W5SNTKH"},"agent_actions":{"view_html":"https://pith.science/pith/2W5SNTKHBIARME5HHRFFN7ATYP","download_json":"https://pith.science/pith/2W5SNTKHBIARME5HHRFFN7ATYP.json","view_paper":"https://pith.science/paper/2W5SNTKH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.11196&json=true","fetch_graph":"https://pith.science/api/pith-number/2W5SNTKHBIARME5HHRFFN7ATYP/graph.json","fetch_events":"https://pith.science/api/pith-number/2W5SNTKHBIARME5HHRFFN7ATYP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2W5SNTKHBIARME5HHRFFN7ATYP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2W5SNTKHBIARME5HHRFFN7ATYP/action/storage_attestation","attest_author":"https://pith.science/pith/2W5SNTKHBIARME5HHRFFN7ATYP/action/author_attestation","sign_citation":"https://pith.science/pith/2W5SNTKHBIARME5HHRFFN7ATYP/action/citation_signature","submit_replication":"https://pith.science/pith/2W5SNTKHBIARME5HHRFFN7ATYP/action/replication_record"}},"created_at":"2026-06-11T00:08:05.838767+00:00","updated_at":"2026-06-11T00:08:05.838767+00:00"}