{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KIZ6EHJI4CJKYIILACXFFUDY5T","short_pith_number":"pith:KIZ6EHJI","schema_version":"1.0","canonical_sha256":"5233e21d28e092ac210b00ae52d078ecc6bcbe7b4cf1ccb612c9a88f4560f7a2","source":{"kind":"arxiv","id":"2608.03689","version":1},"attestation_state":"computed","paper":{"title":"LiveEvalBench: Toward Open-World Evaluation for Web Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.SE"],"primary_cat":"cs.AI","authors_text":"Jun Zhou, Lin Yuan, Wei Chen, Xiaolau Zhang, Yinghao Tang, Yixiao Fu, Yiyao Wang, Zhen Wen","submitted_at":"2026-08-04T13:57:48Z","abstract_excerpt":"Large language models are increasingly capable of synthesizing executable frontend projects, yet existing benchmarks still treat web generation as a static evaluation problem. We argue that frontend artifacts demand a different paradigm: they are interactive rather than static, admit diverse yet equally valid implementations, and evolve faster than rigid pipelines can accommodate. To address these gaps, we present LiveEvalBench, an automated framework that reformulates web-generation evaluation as an agentic, adaptive, and extensible process. LiveEvalBench instantiates evaluation as a collabor"},"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":"2608.03689","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-08-04T13:57:48Z","cross_cats_sorted":["cs.SE"],"title_canon_sha256":"3de76091a6e409e789da6477301e3acd6a88d5af2554c2f4bbb2086716164321","abstract_canon_sha256":"9d6397b40434cbac63fd29dc1f08f751cf0b4992430b812fcb171cde73091083"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-05T01:36:54.016370Z","signature_b64":"piq5KrAAiJtNOOwDNiuClaaSF2HjHL5s1mpNL4709WjBbIvmvWK0B5gN39q/Ba1lNcylDzxarm71SbrO+BktAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5233e21d28e092ac210b00ae52d078ecc6bcbe7b4cf1ccb612c9a88f4560f7a2","last_reissued_at":"2026-08-05T01:36:54.014967Z","signature_status":"signed_v1","first_computed_at":"2026-08-05T01:36:54.014967Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LiveEvalBench: Toward Open-World Evaluation for Web Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.SE"],"primary_cat":"cs.AI","authors_text":"Jun Zhou, Lin Yuan, Wei Chen, Xiaolau Zhang, Yinghao Tang, Yixiao Fu, Yiyao Wang, Zhen Wen","submitted_at":"2026-08-04T13:57:48Z","abstract_excerpt":"Large language models are increasingly capable of synthesizing executable frontend projects, yet existing benchmarks still treat web generation as a static evaluation problem. We argue that frontend artifacts demand a different paradigm: they are interactive rather than static, admit diverse yet equally valid implementations, and evolve faster than rigid pipelines can accommodate. To address these gaps, we present LiveEvalBench, an automated framework that reformulates web-generation evaluation as an agentic, adaptive, and extensible process. LiveEvalBench instantiates evaluation as a collabor"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.03689","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/2608.03689/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":"2608.03689","created_at":"2026-08-05T01:36:54.015426+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.03689v1","created_at":"2026-08-05T01:36:54.015426+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.03689","created_at":"2026-08-05T01:36:54.015426+00:00"},{"alias_kind":"pith_short_12","alias_value":"KIZ6EHJI4CJK","created_at":"2026-08-05T01:36:54.015426+00:00"},{"alias_kind":"pith_short_16","alias_value":"KIZ6EHJI4CJKYIIL","created_at":"2026-08-05T01:36:54.015426+00:00"},{"alias_kind":"pith_short_8","alias_value":"KIZ6EHJI","created_at":"2026-08-05T01:36:54.015426+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/KIZ6EHJI4CJKYIILACXFFUDY5T","json":"https://pith.science/pith/KIZ6EHJI4CJKYIILACXFFUDY5T.json","graph_json":"https://pith.science/api/pith-number/KIZ6EHJI4CJKYIILACXFFUDY5T/graph.json","events_json":"https://pith.science/api/pith-number/KIZ6EHJI4CJKYIILACXFFUDY5T/events.json","paper":"https://pith.science/paper/KIZ6EHJI"},"agent_actions":{"view_html":"https://pith.science/pith/KIZ6EHJI4CJKYIILACXFFUDY5T","download_json":"https://pith.science/pith/KIZ6EHJI4CJKYIILACXFFUDY5T.json","view_paper":"https://pith.science/paper/KIZ6EHJI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.03689&json=true","fetch_graph":"https://pith.science/api/pith-number/KIZ6EHJI4CJKYIILACXFFUDY5T/graph.json","fetch_events":"https://pith.science/api/pith-number/KIZ6EHJI4CJKYIILACXFFUDY5T/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KIZ6EHJI4CJKYIILACXFFUDY5T/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KIZ6EHJI4CJKYIILACXFFUDY5T/action/storage_attestation","attest_author":"https://pith.science/pith/KIZ6EHJI4CJKYIILACXFFUDY5T/action/author_attestation","sign_citation":"https://pith.science/pith/KIZ6EHJI4CJKYIILACXFFUDY5T/action/citation_signature","submit_replication":"https://pith.science/pith/KIZ6EHJI4CJKYIILACXFFUDY5T/action/replication_record"}},"created_at":"2026-08-05T01:36:54.015426+00:00","updated_at":"2026-08-05T01:36:54.015426+00:00"}