{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:DQ62XXYNEE4ME3YAHIJTK6X3SU","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"30619f426ef8d9d7aa732d642b2477ab66d30f5a9cfeed9154c23766866db1f1","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.SD","submitted_at":"2026-06-07T16:32:59Z","title_canon_sha256":"b59eef23bb88bde8991a103838549513d8496240be15900954cee755c6047848"},"schema_version":"1.0","source":{"id":"2606.08722","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.08722","created_at":"2026-06-09T02:07:35Z"},{"alias_kind":"arxiv_version","alias_value":"2606.08722v1","created_at":"2026-06-09T02:07:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.08722","created_at":"2026-06-09T02:07:35Z"},{"alias_kind":"pith_short_12","alias_value":"DQ62XXYNEE4M","created_at":"2026-06-09T02:07:35Z"},{"alias_kind":"pith_short_16","alias_value":"DQ62XXYNEE4ME3YA","created_at":"2026-06-09T02:07:35Z"},{"alias_kind":"pith_short_8","alias_value":"DQ62XXYN","created_at":"2026-06-09T02:07:35Z"}],"graph_snapshots":[{"event_id":"sha256:794cc7a097d6a1a4364e6a73071fb553ee2d53a0f23618dcb6d60be528fe9bd0","target":"graph","created_at":"2026-06-09T02:07:35Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2606.08722/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Symbolic music evaluation for large language models remains fragmented across representations, datasets, and metrics. We introduce LilyBench, a LilyPond-based benchmark that jointly evaluates symbolic music generation and music understanding on the same family of open-weight LLMs. The benchmark includes a 200-prompt generation suite and ten understanding tasks adapted from ABC-Eval, covering syntax, metadata prediction, structural sequencing, and music recognition. Generation quality is evaluated using compile rate, MusPy descriptor distributions via Jensen-Shannon similarity, and LilyBERT-bas","authors_text":"Andrea Poltronieri, Antonio Rod\\`a, Matteo Spanio, Mohammad Torabi","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.SD","submitted_at":"2026-06-07T16:32:59Z","title":"Can LLMs understand LilyPond? A benchmark for symbolic music generation and understanding"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.08722","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:81cb5468476e6e47eec56ce0d4c17c7333b27d04ef06922b1b3db201acb20952","target":"record","created_at":"2026-06-09T02:07:35Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"30619f426ef8d9d7aa732d642b2477ab66d30f5a9cfeed9154c23766866db1f1","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.SD","submitted_at":"2026-06-07T16:32:59Z","title_canon_sha256":"b59eef23bb88bde8991a103838549513d8496240be15900954cee755c6047848"},"schema_version":"1.0","source":{"id":"2606.08722","kind":"arxiv","version":1}},"canonical_sha256":"1c3dabdf0d2138c26f003a13357afb953ea6cb8918665c6e22fa1de0af456dcb","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1c3dabdf0d2138c26f003a13357afb953ea6cb8918665c6e22fa1de0af456dcb","first_computed_at":"2026-06-09T02:07:35.378799Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-09T02:07:35.378799Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"YwDvf7PN6A6LgdPeAR5CbncNCwdm3ArOgcsnaMfsxgrDBmOWq08bR8aw5POzYSw/dTNyT1NXAKQPaIpKhvN7AA==","signature_status":"signed_v1","signed_at":"2026-06-09T02:07:35.379483Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.08722","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:81cb5468476e6e47eec56ce0d4c17c7333b27d04ef06922b1b3db201acb20952","sha256:794cc7a097d6a1a4364e6a73071fb553ee2d53a0f23618dcb6d60be528fe9bd0"],"state_sha256":"bba054be501bf1455a76e68210b43753e6b4bf3a4ad020942a0210810a1b4e0c"}