{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:7MHR6EYM7DUZHWO5KDWZIUCZI4","short_pith_number":"pith:7MHR6EYM","schema_version":"1.0","canonical_sha256":"fb0f1f130cf8e993d9dd50ed94505947087b1e7881e3950bffae1ac895b647b0","source":{"kind":"arxiv","id":"2608.10567","version":1},"attestation_state":"computed","paper":{"title":"DashArena: Benchmarking LLMs on Interactive Analytic Dashboard Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Dazhen Deng, Xiaotong Wang","submitted_at":"2026-08-11T06:54:37Z","abstract_excerpt":"Analytic dashboards combine coordinated views and interactions for data exploration and decision-making. Recent models can generate them from data and natural-language goals, but evaluating their usefulness remains difficult. Dashboard generation is open-ended, and neither static appearance nor successful execution alone captures analytical support and interaction quality. We introduce DashArena, to our knowledge the first benchmark for open-ended, task-grounded generation of interactive analytic dashboards. Its key innovation is to require each system to generate both a dashboard and a replay"},"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":"2608.10567","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-08-11T06:54:37Z","cross_cats_sorted":[],"title_canon_sha256":"0ba37626e91becf073f23c4bf3d270196e575dcb10c14a97713c9adc2259729c","abstract_canon_sha256":"343ac33db842f6d7b29ad77bae60f18a3c12c150a00bb407e1b6bb6b6cdafeb4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-12T01:22:46.351093Z","signature_b64":"WzutDrfdt6A1UWJBiMVxRQ3kdXJio8LWTAMCisCwbgyK409r4dcDg8cWCqmPHh3PUUGsoG2SiMsgQpxrdiwDDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fb0f1f130cf8e993d9dd50ed94505947087b1e7881e3950bffae1ac895b647b0","last_reissued_at":"2026-08-12T01:22:46.349562Z","signature_status":"signed_v1","first_computed_at":"2026-08-12T01:22:46.349562Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DashArena: Benchmarking LLMs on Interactive Analytic Dashboard Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Dazhen Deng, Xiaotong Wang","submitted_at":"2026-08-11T06:54:37Z","abstract_excerpt":"Analytic dashboards combine coordinated views and interactions for data exploration and decision-making. Recent models can generate them from data and natural-language goals, but evaluating their usefulness remains difficult. Dashboard generation is open-ended, and neither static appearance nor successful execution alone captures analytical support and interaction quality. We introduce DashArena, to our knowledge the first benchmark for open-ended, task-grounded generation of interactive analytic dashboards. Its key innovation is to require each system to generate both a dashboard and a replay"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.10567","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/2608.10567/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":"2608.10567","created_at":"2026-08-12T01:22:46.353342+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.10567v1","created_at":"2026-08-12T01:22:46.353342+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.10567","created_at":"2026-08-12T01:22:46.353342+00:00"},{"alias_kind":"pith_short_12","alias_value":"7MHR6EYM7DUZ","created_at":"2026-08-12T01:22:46.353342+00:00"},{"alias_kind":"pith_short_16","alias_value":"7MHR6EYM7DUZHWO5","created_at":"2026-08-12T01:22:46.353342+00:00"},{"alias_kind":"pith_short_8","alias_value":"7MHR6EYM","created_at":"2026-08-12T01:22:46.353342+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/7MHR6EYM7DUZHWO5KDWZIUCZI4","json":"https://pith.science/pith/7MHR6EYM7DUZHWO5KDWZIUCZI4.json","graph_json":"https://pith.science/api/pith-number/7MHR6EYM7DUZHWO5KDWZIUCZI4/graph.json","events_json":"https://pith.science/api/pith-number/7MHR6EYM7DUZHWO5KDWZIUCZI4/events.json","paper":"https://pith.science/paper/7MHR6EYM"},"agent_actions":{"view_html":"https://pith.science/pith/7MHR6EYM7DUZHWO5KDWZIUCZI4","download_json":"https://pith.science/pith/7MHR6EYM7DUZHWO5KDWZIUCZI4.json","view_paper":"https://pith.science/paper/7MHR6EYM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.10567&json=true","fetch_graph":"https://pith.science/api/pith-number/7MHR6EYM7DUZHWO5KDWZIUCZI4/graph.json","fetch_events":"https://pith.science/api/pith-number/7MHR6EYM7DUZHWO5KDWZIUCZI4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7MHR6EYM7DUZHWO5KDWZIUCZI4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7MHR6EYM7DUZHWO5KDWZIUCZI4/action/storage_attestation","attest_author":"https://pith.science/pith/7MHR6EYM7DUZHWO5KDWZIUCZI4/action/author_attestation","sign_citation":"https://pith.science/pith/7MHR6EYM7DUZHWO5KDWZIUCZI4/action/citation_signature","submit_replication":"https://pith.science/pith/7MHR6EYM7DUZHWO5KDWZIUCZI4/action/replication_record"}},"created_at":"2026-08-12T01:22:46.353342+00:00","updated_at":"2026-08-12T01:22:46.353342+00:00"}