{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NCB355APP2FBL6EXUJGNO5C75D","short_pith_number":"pith:NCB355AP","schema_version":"1.0","canonical_sha256":"6883bef40f7e8a15f897a24cd7745fe8dc124f5181db4dd1178f1bac5e3f5d01","source":{"kind":"arxiv","id":"2605.19394","version":1},"attestation_state":"computed","paper":{"title":"EmbGen: Teaching with Reassembled Corpora","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Andrea Nicastro, Anna Leontjeva, Arun K Lenin, Kai Rouse","submitted_at":"2026-05-19T05:40:12Z","abstract_excerpt":"Adapting small instruction-tuned models to specialized domains often relies on supervised fine-tuning (SFT) on curated instruction-response examples, which is expensive to collect at scale. Synthetic training examples generated by a teacher LLM from a domain corpus can reduce this cost, but existing pipelines can produce homogenized outputs and do not consistently capture cross-passage or cross-document dependencies. We introduce EmbGen, a synthetic data generation pipeline that decomposes a corpus into entity-description pairs, reassembles them using semantic structure inferred from embedding"},"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":"2605.19394","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-19T05:40:12Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ee34a0714ae02233056bf3f0dd0356d8d9e7ad6303ae81694c8115ba10530a60","abstract_canon_sha256":"e86b514e222667e41bb177c83c88f69307cd77e6f7b7568eb85a90000500ad74"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T01:05:43.496688Z","signature_b64":"iS1sRVTbxrfSydfgnvTTB+yqPgZbENnEirTq6TbknBuUVcZ5VLgyRMLovLQciBKOM/v3wyD5A8LWV/53axofDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6883bef40f7e8a15f897a24cd7745fe8dc124f5181db4dd1178f1bac5e3f5d01","last_reissued_at":"2026-05-20T01:05:43.495825Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T01:05:43.495825Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"EmbGen: Teaching with Reassembled Corpora","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Andrea Nicastro, Anna Leontjeva, Arun K Lenin, Kai Rouse","submitted_at":"2026-05-19T05:40:12Z","abstract_excerpt":"Adapting small instruction-tuned models to specialized domains often relies on supervised fine-tuning (SFT) on curated instruction-response examples, which is expensive to collect at scale. Synthetic training examples generated by a teacher LLM from a domain corpus can reduce this cost, but existing pipelines can produce homogenized outputs and do not consistently capture cross-passage or cross-document dependencies. We introduce EmbGen, a synthetic data generation pipeline that decomposes a corpus into entity-description pairs, reassembles them using semantic structure inferred from embedding"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.19394","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/2605.19394/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":"2605.19394","created_at":"2026-05-20T01:05:43.495972+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.19394v1","created_at":"2026-05-20T01:05:43.495972+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.19394","created_at":"2026-05-20T01:05:43.495972+00:00"},{"alias_kind":"pith_short_12","alias_value":"NCB355APP2FB","created_at":"2026-05-20T01:05:43.495972+00:00"},{"alias_kind":"pith_short_16","alias_value":"NCB355APP2FBL6EX","created_at":"2026-05-20T01:05:43.495972+00:00"},{"alias_kind":"pith_short_8","alias_value":"NCB355AP","created_at":"2026-05-20T01:05:43.495972+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/NCB355APP2FBL6EXUJGNO5C75D","json":"https://pith.science/pith/NCB355APP2FBL6EXUJGNO5C75D.json","graph_json":"https://pith.science/api/pith-number/NCB355APP2FBL6EXUJGNO5C75D/graph.json","events_json":"https://pith.science/api/pith-number/NCB355APP2FBL6EXUJGNO5C75D/events.json","paper":"https://pith.science/paper/NCB355AP"},"agent_actions":{"view_html":"https://pith.science/pith/NCB355APP2FBL6EXUJGNO5C75D","download_json":"https://pith.science/pith/NCB355APP2FBL6EXUJGNO5C75D.json","view_paper":"https://pith.science/paper/NCB355AP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.19394&json=true","fetch_graph":"https://pith.science/api/pith-number/NCB355APP2FBL6EXUJGNO5C75D/graph.json","fetch_events":"https://pith.science/api/pith-number/NCB355APP2FBL6EXUJGNO5C75D/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NCB355APP2FBL6EXUJGNO5C75D/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NCB355APP2FBL6EXUJGNO5C75D/action/storage_attestation","attest_author":"https://pith.science/pith/NCB355APP2FBL6EXUJGNO5C75D/action/author_attestation","sign_citation":"https://pith.science/pith/NCB355APP2FBL6EXUJGNO5C75D/action/citation_signature","submit_replication":"https://pith.science/pith/NCB355APP2FBL6EXUJGNO5C75D/action/replication_record"}},"created_at":"2026-05-20T01:05:43.495972+00:00","updated_at":"2026-05-20T01:05:43.495972+00:00"}