{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:EAHKA6JEFUCTPRFO62S3574NI6","short_pith_number":"pith:EAHKA6JE","schema_version":"1.0","canonical_sha256":"200ea079242d0537c4aef6a5beff8d478cada9f6fa81a93970c546b7b8c8153b","source":{"kind":"arxiv","id":"2501.08648","version":2},"attestation_state":"computed","paper":{"title":"MAGNET: Augmenting Generative Decoders with Representation Learning and Infilling Capabilities","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aditi Tiwari, Handong Zhao, Jing Shi, John Collomosse, Kushal Kafle, Savya Khosla, Simon Jenni","submitted_at":"2025-01-15T08:24:03Z","abstract_excerpt":"While originally designed for unidirectional generative modeling, decoder-only large language models (LLMs) are increasingly being adapted for bidirectional modeling. However, unidirectional and bidirectional models are typically trained separately with distinct objectives (generation and representation learning). This separation overlooks the opportunity for developing a more versatile language model and for these objectives to complement each other. In this work, we propose MAGNET, a method for adapting decoder-only LLMs to generate robust representations and infill missing text spans. MAGNE"},"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":"2501.08648","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-01-15T08:24:03Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"9c9f4bd2b4b2a41952b541c18d7bc2fa0d6e9591bccb9fca0c02b7de3b27c988","abstract_canon_sha256":"c84564d6b3ad7e6c3e1f93191ddab298b160cc7800a6753fdb872ab7166887fc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:14:08.070422Z","signature_b64":"vpdLEKmS9ITbez492CFmHLPaUZhzgy802zjwqIlCp5ptRwU1ZKGa6v/u5R2x/FfiomvhrurcLb1WW/z7jCLyAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"200ea079242d0537c4aef6a5beff8d478cada9f6fa81a93970c546b7b8c8153b","last_reissued_at":"2026-07-05T10:14:08.069881Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:14:08.069881Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MAGNET: Augmenting Generative Decoders with Representation Learning and Infilling Capabilities","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aditi Tiwari, Handong Zhao, Jing Shi, John Collomosse, Kushal Kafle, Savya Khosla, Simon Jenni","submitted_at":"2025-01-15T08:24:03Z","abstract_excerpt":"While originally designed for unidirectional generative modeling, decoder-only large language models (LLMs) are increasingly being adapted for bidirectional modeling. However, unidirectional and bidirectional models are typically trained separately with distinct objectives (generation and representation learning). This separation overlooks the opportunity for developing a more versatile language model and for these objectives to complement each other. In this work, we propose MAGNET, a method for adapting decoder-only LLMs to generate robust representations and infill missing text spans. MAGNE"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.08648","kind":"arxiv","version":2},"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/2501.08648/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":"2501.08648","created_at":"2026-07-05T10:14:08.069950+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.08648v2","created_at":"2026-07-05T10:14:08.069950+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.08648","created_at":"2026-07-05T10:14:08.069950+00:00"},{"alias_kind":"pith_short_12","alias_value":"EAHKA6JEFUCT","created_at":"2026-07-05T10:14:08.069950+00:00"},{"alias_kind":"pith_short_16","alias_value":"EAHKA6JEFUCTPRFO","created_at":"2026-07-05T10:14:08.069950+00:00"},{"alias_kind":"pith_short_8","alias_value":"EAHKA6JE","created_at":"2026-07-05T10:14:08.069950+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.06622","citing_title":"FuDoBa: Fusing Document and Knowledge Graph-based Representations with Bayesian Optimisation","ref_index":18,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EAHKA6JEFUCTPRFO62S3574NI6","json":"https://pith.science/pith/EAHKA6JEFUCTPRFO62S3574NI6.json","graph_json":"https://pith.science/api/pith-number/EAHKA6JEFUCTPRFO62S3574NI6/graph.json","events_json":"https://pith.science/api/pith-number/EAHKA6JEFUCTPRFO62S3574NI6/events.json","paper":"https://pith.science/paper/EAHKA6JE"},"agent_actions":{"view_html":"https://pith.science/pith/EAHKA6JEFUCTPRFO62S3574NI6","download_json":"https://pith.science/pith/EAHKA6JEFUCTPRFO62S3574NI6.json","view_paper":"https://pith.science/paper/EAHKA6JE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.08648&json=true","fetch_graph":"https://pith.science/api/pith-number/EAHKA6JEFUCTPRFO62S3574NI6/graph.json","fetch_events":"https://pith.science/api/pith-number/EAHKA6JEFUCTPRFO62S3574NI6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EAHKA6JEFUCTPRFO62S3574NI6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EAHKA6JEFUCTPRFO62S3574NI6/action/storage_attestation","attest_author":"https://pith.science/pith/EAHKA6JEFUCTPRFO62S3574NI6/action/author_attestation","sign_citation":"https://pith.science/pith/EAHKA6JEFUCTPRFO62S3574NI6/action/citation_signature","submit_replication":"https://pith.science/pith/EAHKA6JEFUCTPRFO62S3574NI6/action/replication_record"}},"created_at":"2026-07-05T10:14:08.069950+00:00","updated_at":"2026-07-05T10:14:08.069950+00:00"}