{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:SNDHU6XKKHJCX5UYQUNOMO5VQT","short_pith_number":"pith:SNDHU6XK","schema_version":"1.0","canonical_sha256":"93467a7aea51d22bf698851ae63bb584e3b349f8a27ea9ca0d9664232b23763f","source":{"kind":"arxiv","id":"2507.15236","version":1},"attestation_state":"computed","paper":{"title":"SOI Matters: Analyzing Multi-Setting Training Dynamics in Pretrained Language Models via Subsets of Interest","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Amirhossein Dabiriaghdam, Mohammadreza Bakhtiari, Shayan Vassef, Yadollah Yaghoobzadeh","submitted_at":"2025-07-21T04:43:21Z","abstract_excerpt":"This work investigates the impact of multi-task, multi-lingual, and multi-source learning approaches on the robustness and performance of pretrained language models. To enhance this analysis, we introduce Subsets of Interest (SOI), a novel categorization framework that identifies six distinct learning behavior patterns during training, including forgettable examples, unlearned examples, and always correct examples. Through SOI transition heatmaps and dataset cartography visualization, we analyze how examples shift between these categories when transitioning from single-setting to multi-setting"},"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":"2507.15236","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-21T04:43:21Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"a189de23a6c7918a62b25b2719a6e60f70f3d641ccdf15d6f72f0316cf5ba7ec","abstract_canon_sha256":"7db41581b632845a8ffcce5461cc6ae7253a9bd47345a00e2e4fda25ac9cb7f1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:40:20.764944Z","signature_b64":"dap+uCvesL4emSArUDhomIR+3K+XQRa+PV16mANXedUzZdstmRyFLtY0v3ayJKxf+zY5+zBmATcmXuyn3XNYBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"93467a7aea51d22bf698851ae63bb584e3b349f8a27ea9ca0d9664232b23763f","last_reissued_at":"2026-07-05T11:40:20.764446Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:40:20.764446Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SOI Matters: Analyzing Multi-Setting Training Dynamics in Pretrained Language Models via Subsets of Interest","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Amirhossein Dabiriaghdam, Mohammadreza Bakhtiari, Shayan Vassef, Yadollah Yaghoobzadeh","submitted_at":"2025-07-21T04:43:21Z","abstract_excerpt":"This work investigates the impact of multi-task, multi-lingual, and multi-source learning approaches on the robustness and performance of pretrained language models. To enhance this analysis, we introduce Subsets of Interest (SOI), a novel categorization framework that identifies six distinct learning behavior patterns during training, including forgettable examples, unlearned examples, and always correct examples. Through SOI transition heatmaps and dataset cartography visualization, we analyze how examples shift between these categories when transitioning from single-setting to multi-setting"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.15236","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/2507.15236/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":"2507.15236","created_at":"2026-07-05T11:40:20.764504+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.15236v1","created_at":"2026-07-05T11:40:20.764504+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.15236","created_at":"2026-07-05T11:40:20.764504+00:00"},{"alias_kind":"pith_short_12","alias_value":"SNDHU6XKKHJC","created_at":"2026-07-05T11:40:20.764504+00:00"},{"alias_kind":"pith_short_16","alias_value":"SNDHU6XKKHJCX5UY","created_at":"2026-07-05T11:40:20.764504+00:00"},{"alias_kind":"pith_short_8","alias_value":"SNDHU6XK","created_at":"2026-07-05T11:40:20.764504+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/SNDHU6XKKHJCX5UYQUNOMO5VQT","json":"https://pith.science/pith/SNDHU6XKKHJCX5UYQUNOMO5VQT.json","graph_json":"https://pith.science/api/pith-number/SNDHU6XKKHJCX5UYQUNOMO5VQT/graph.json","events_json":"https://pith.science/api/pith-number/SNDHU6XKKHJCX5UYQUNOMO5VQT/events.json","paper":"https://pith.science/paper/SNDHU6XK"},"agent_actions":{"view_html":"https://pith.science/pith/SNDHU6XKKHJCX5UYQUNOMO5VQT","download_json":"https://pith.science/pith/SNDHU6XKKHJCX5UYQUNOMO5VQT.json","view_paper":"https://pith.science/paper/SNDHU6XK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.15236&json=true","fetch_graph":"https://pith.science/api/pith-number/SNDHU6XKKHJCX5UYQUNOMO5VQT/graph.json","fetch_events":"https://pith.science/api/pith-number/SNDHU6XKKHJCX5UYQUNOMO5VQT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SNDHU6XKKHJCX5UYQUNOMO5VQT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SNDHU6XKKHJCX5UYQUNOMO5VQT/action/storage_attestation","attest_author":"https://pith.science/pith/SNDHU6XKKHJCX5UYQUNOMO5VQT/action/author_attestation","sign_citation":"https://pith.science/pith/SNDHU6XKKHJCX5UYQUNOMO5VQT/action/citation_signature","submit_replication":"https://pith.science/pith/SNDHU6XKKHJCX5UYQUNOMO5VQT/action/replication_record"}},"created_at":"2026-07-05T11:40:20.764504+00:00","updated_at":"2026-07-05T11:40:20.764504+00:00"}