{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:FUGD3Y2UVVEJMLF43LK7WZO3C6","short_pith_number":"pith:FUGD3Y2U","canonical_record":{"source":{"id":"2407.18887","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-07-26T17:36:40Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"175b4b5a00f31b3fa8869f220529d23a820a6966f950050277ad6f01d79a22de","abstract_canon_sha256":"b2c3aef19772f7351dbbabc31a7f0bae8a3d30c9b24fad08d1d1a253b18bbfa8"},"schema_version":"1.0"},"canonical_sha256":"2d0c3de354ad48962cbcdad5fb65db17831bb385d6d280a1b05512ec2773258a","source":{"kind":"arxiv","id":"2407.18887","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.18887","created_at":"2026-07-05T08:48:55Z"},{"alias_kind":"arxiv_version","alias_value":"2407.18887v1","created_at":"2026-07-05T08:48:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.18887","created_at":"2026-07-05T08:48:55Z"},{"alias_kind":"pith_short_12","alias_value":"FUGD3Y2UVVEJ","created_at":"2026-07-05T08:48:55Z"},{"alias_kind":"pith_short_16","alias_value":"FUGD3Y2UVVEJMLF4","created_at":"2026-07-05T08:48:55Z"},{"alias_kind":"pith_short_8","alias_value":"FUGD3Y2U","created_at":"2026-07-05T08:48:55Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:FUGD3Y2UVVEJMLF43LK7WZO3C6","target":"record","payload":{"canonical_record":{"source":{"id":"2407.18887","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-07-26T17:36:40Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"175b4b5a00f31b3fa8869f220529d23a820a6966f950050277ad6f01d79a22de","abstract_canon_sha256":"b2c3aef19772f7351dbbabc31a7f0bae8a3d30c9b24fad08d1d1a253b18bbfa8"},"schema_version":"1.0"},"canonical_sha256":"2d0c3de354ad48962cbcdad5fb65db17831bb385d6d280a1b05512ec2773258a","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:48:55.882246Z","signature_b64":"YIHNSAvO+LC40B9fhQMLyFTJBKLwLcU8naS0Jl5S4smfbrY8n9/gU7qC1O97sqplaxhbSYNMTEt78lED0DHVAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2d0c3de354ad48962cbcdad5fb65db17831bb385d6d280a1b05512ec2773258a","last_reissued_at":"2026-07-05T08:48:55.881834Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:48:55.881834Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2407.18887","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:48:55Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"pB5o9q73z7FmSOh1PXrVWEsSboH8XkScomF/mE0s4Nr0gQv9qo5cT6RC7I8dwaNjzs5r3OxJO9n9e69poiRcCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T11:24:21.265877Z"},"content_sha256":"032f13483391e0b8ae04f0a696f909676b2c6d975c2138b8c6a8263dff729138","schema_version":"1.0","event_id":"sha256:032f13483391e0b8ae04f0a696f909676b2c6d975c2138b8c6a8263dff729138"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:FUGD3Y2UVVEJMLF43LK7WZO3C6","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Embedding And Clustering Your Data Can Improve Contrastive Pretraining","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Luke Merrick","submitted_at":"2024-07-26T17:36:40Z","abstract_excerpt":"Recent studies of large-scale contrastive pretraining in the text embedding domain show that using single-source minibatches, rather than mixed-source minibatches, can substantially improve overall model accuracy. In this work, we explore extending training data stratification beyond source granularity by leveraging a pretrained text embedding model and the classic k-means clustering algorithm to further split training data apart by the semantic clusters within each source. Experimentally, we observe a notable increase in NDCG@10 when pretraining a BERT-based text embedding model on query-pass"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.18887","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/2407.18887/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:48:55Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SWR3uFqHm8D7W8tN49GaaY9hvf45CCmg7QB+qdvXMYxT9/vGJ0pbJLkNbEPy9uZR/Rdq511++fsJYpCe10IUDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T11:24:21.266401Z"},"content_sha256":"b865956e0505c4cbb46d707a8313142e5fb068bde359d52b478e04e3f7cd262e","schema_version":"1.0","event_id":"sha256:b865956e0505c4cbb46d707a8313142e5fb068bde359d52b478e04e3f7cd262e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/FUGD3Y2UVVEJMLF43LK7WZO3C6/bundle.json","state_url":"https://pith.science/pith/FUGD3Y2UVVEJMLF43LK7WZO3C6/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/FUGD3Y2UVVEJMLF43LK7WZO3C6/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-08T11:24:21Z","links":{"resolver":"https://pith.science/pith/FUGD3Y2UVVEJMLF43LK7WZO3C6","bundle":"https://pith.science/pith/FUGD3Y2UVVEJMLF43LK7WZO3C6/bundle.json","state":"https://pith.science/pith/FUGD3Y2UVVEJMLF43LK7WZO3C6/state.json","well_known_bundle":"https://pith.science/.well-known/pith/FUGD3Y2UVVEJMLF43LK7WZO3C6/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:FUGD3Y2UVVEJMLF43LK7WZO3C6","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":"b2c3aef19772f7351dbbabc31a7f0bae8a3d30c9b24fad08d1d1a253b18bbfa8","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-07-26T17:36:40Z","title_canon_sha256":"175b4b5a00f31b3fa8869f220529d23a820a6966f950050277ad6f01d79a22de"},"schema_version":"1.0","source":{"id":"2407.18887","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2407.18887","created_at":"2026-07-05T08:48:55Z"},{"alias_kind":"arxiv_version","alias_value":"2407.18887v1","created_at":"2026-07-05T08:48:55Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.18887","created_at":"2026-07-05T08:48:55Z"},{"alias_kind":"pith_short_12","alias_value":"FUGD3Y2UVVEJ","created_at":"2026-07-05T08:48:55Z"},{"alias_kind":"pith_short_16","alias_value":"FUGD3Y2UVVEJMLF4","created_at":"2026-07-05T08:48:55Z"},{"alias_kind":"pith_short_8","alias_value":"FUGD3Y2U","created_at":"2026-07-05T08:48:55Z"}],"graph_snapshots":[{"event_id":"sha256:b865956e0505c4cbb46d707a8313142e5fb068bde359d52b478e04e3f7cd262e","target":"graph","created_at":"2026-07-05T08:48:55Z","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/2407.18887/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent studies of large-scale contrastive pretraining in the text embedding domain show that using single-source minibatches, rather than mixed-source minibatches, can substantially improve overall model accuracy. In this work, we explore extending training data stratification beyond source granularity by leveraging a pretrained text embedding model and the classic k-means clustering algorithm to further split training data apart by the semantic clusters within each source. Experimentally, we observe a notable increase in NDCG@10 when pretraining a BERT-based text embedding model on query-pass","authors_text":"Luke Merrick","cross_cats":["cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-07-26T17:36:40Z","title":"Embedding And Clustering Your Data Can Improve Contrastive Pretraining"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.18887","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":"sha256:032f13483391e0b8ae04f0a696f909676b2c6d975c2138b8c6a8263dff729138","target":"record","created_at":"2026-07-05T08:48:55Z","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":"b2c3aef19772f7351dbbabc31a7f0bae8a3d30c9b24fad08d1d1a253b18bbfa8","cross_cats_sorted":["cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-07-26T17:36:40Z","title_canon_sha256":"175b4b5a00f31b3fa8869f220529d23a820a6966f950050277ad6f01d79a22de"},"schema_version":"1.0","source":{"id":"2407.18887","kind":"arxiv","version":1}},"canonical_sha256":"2d0c3de354ad48962cbcdad5fb65db17831bb385d6d280a1b05512ec2773258a","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2d0c3de354ad48962cbcdad5fb65db17831bb385d6d280a1b05512ec2773258a","first_computed_at":"2026-07-05T08:48:55.881834Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:48:55.881834Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"YIHNSAvO+LC40B9fhQMLyFTJBKLwLcU8naS0Jl5S4smfbrY8n9/gU7qC1O97sqplaxhbSYNMTEt78lED0DHVAg==","signature_status":"signed_v1","signed_at":"2026-07-05T08:48:55.882246Z","signed_message":"canonical_sha256_bytes"},"source_id":"2407.18887","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:032f13483391e0b8ae04f0a696f909676b2c6d975c2138b8c6a8263dff729138","sha256:b865956e0505c4cbb46d707a8313142e5fb068bde359d52b478e04e3f7cd262e"],"state_sha256":"d0c9fca3db39b45caeafa6a8e7c18ee4d9d257bc7650e243fd446d3d17272085"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2rQmLr8nGZ0OlnxrayXqE+ms6mK8+2pJjwrM7kOvHJ+8unmqCvQblMGWi3HNixb64rGalzZBoPQ9ZSSz4+fjCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T11:24:21.270472Z","bundle_sha256":"3c3ce3f63b4dabbff864c7bb86435c9b23d083da658cba4bad9ccb654a3ea243"}}