{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5FFK5MUPL32NPC5C4P42AIETNX","short_pith_number":"pith:5FFK5MUP","schema_version":"1.0","canonical_sha256":"e94aaeb28f5ef4d78ba2e3f9a020936dc2e298c0ab2e840e54afb9e6df3a038b","source":{"kind":"arxiv","id":"2506.20917","version":1},"attestation_state":"computed","paper":{"title":"Optimising Language Models for Downstream Tasks: A Post-Training Perspective","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Zhengyan Shi","submitted_at":"2025-06-26T00:49:35Z","abstract_excerpt":"Language models (LMs) have demonstrated remarkable capabilities in NLP, yet adapting them efficiently and robustly to specific tasks remains challenging. As their scale and complexity grow, fine-tuning LMs on labelled data often underutilizes available unlabelled data, leads to overfitting on small task-specific sets, and imposes significant computational costs. These limitations hamper their application to the open-ended landscape of real-world language tasks.\n  This thesis proposes a series of methods to better adapt LMs to downstream applications. First, we explore strategies for extracting"},"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":"2506.20917","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-06-26T00:49:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"47c07d6e38b0c8d80462da96ab717e69708750b9f89b8213275ea91f0e73d63a","abstract_canon_sha256":"216eee09c3ab8fc3791280059dcef504cd007d66a063ce35b47a8385657d14c1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:27:36.079901Z","signature_b64":"R2hRROzh/cCFgo2d9YyoSEedE2Bh4ysG9kyZL0D/HNgPcaKfh494b/Ufov7oVAzWWWN5eKUBzUWF6hwtJ5P4Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e94aaeb28f5ef4d78ba2e3f9a020936dc2e298c0ab2e840e54afb9e6df3a038b","last_reissued_at":"2026-07-05T11:27:36.079509Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:27:36.079509Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Optimising Language Models for Downstream Tasks: A Post-Training Perspective","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Zhengyan Shi","submitted_at":"2025-06-26T00:49:35Z","abstract_excerpt":"Language models (LMs) have demonstrated remarkable capabilities in NLP, yet adapting them efficiently and robustly to specific tasks remains challenging. As their scale and complexity grow, fine-tuning LMs on labelled data often underutilizes available unlabelled data, leads to overfitting on small task-specific sets, and imposes significant computational costs. These limitations hamper their application to the open-ended landscape of real-world language tasks.\n  This thesis proposes a series of methods to better adapt LMs to downstream applications. First, we explore strategies for extracting"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.20917","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/2506.20917/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":"2506.20917","created_at":"2026-07-05T11:27:36.079561+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.20917v1","created_at":"2026-07-05T11:27:36.079561+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.20917","created_at":"2026-07-05T11:27:36.079561+00:00"},{"alias_kind":"pith_short_12","alias_value":"5FFK5MUPL32N","created_at":"2026-07-05T11:27:36.079561+00:00"},{"alias_kind":"pith_short_16","alias_value":"5FFK5MUPL32NPC5C","created_at":"2026-07-05T11:27:36.079561+00:00"},{"alias_kind":"pith_short_8","alias_value":"5FFK5MUP","created_at":"2026-07-05T11:27:36.079561+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/5FFK5MUPL32NPC5C4P42AIETNX","json":"https://pith.science/pith/5FFK5MUPL32NPC5C4P42AIETNX.json","graph_json":"https://pith.science/api/pith-number/5FFK5MUPL32NPC5C4P42AIETNX/graph.json","events_json":"https://pith.science/api/pith-number/5FFK5MUPL32NPC5C4P42AIETNX/events.json","paper":"https://pith.science/paper/5FFK5MUP"},"agent_actions":{"view_html":"https://pith.science/pith/5FFK5MUPL32NPC5C4P42AIETNX","download_json":"https://pith.science/pith/5FFK5MUPL32NPC5C4P42AIETNX.json","view_paper":"https://pith.science/paper/5FFK5MUP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.20917&json=true","fetch_graph":"https://pith.science/api/pith-number/5FFK5MUPL32NPC5C4P42AIETNX/graph.json","fetch_events":"https://pith.science/api/pith-number/5FFK5MUPL32NPC5C4P42AIETNX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5FFK5MUPL32NPC5C4P42AIETNX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5FFK5MUPL32NPC5C4P42AIETNX/action/storage_attestation","attest_author":"https://pith.science/pith/5FFK5MUPL32NPC5C4P42AIETNX/action/author_attestation","sign_citation":"https://pith.science/pith/5FFK5MUPL32NPC5C4P42AIETNX/action/citation_signature","submit_replication":"https://pith.science/pith/5FFK5MUPL32NPC5C4P42AIETNX/action/replication_record"}},"created_at":"2026-07-05T11:27:36.079561+00:00","updated_at":"2026-07-05T11:27:36.079561+00:00"}