{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:E6OAEVQZCZ724XZVXJ4EYYSPOW","short_pith_number":"pith:E6OAEVQZ","schema_version":"1.0","canonical_sha256":"279c025619167fae5f35ba784c624f75bc6dc98fb28d0e5c48d824f0976f100b","source":{"kind":"arxiv","id":"2311.01732","version":2},"attestation_state":"computed","paper":{"title":"Proto-lm: A Prototypical Network-Based Framework for Built-in Interpretability in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Saeed Hassanpour, Sean Xie, Soroush Vosoughi","submitted_at":"2023-11-03T05:55:32Z","abstract_excerpt":"Large Language Models (LLMs) have significantly advanced the field of Natural Language Processing (NLP), but their lack of interpretability has been a major concern. Current methods for interpreting LLMs are post hoc, applied after inference time, and have limitations such as their focus on low-level features and lack of explainability at higher level text units. In this work, we introduce proto-lm, a prototypical network-based white-box framework that allows LLMs to learn immediately interpretable embeddings during the fine-tuning stage while maintaining competitive performance. Our method's "},"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":"2311.01732","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-03T05:55:32Z","cross_cats_sorted":[],"title_canon_sha256":"a5d52e28b108fa34ff9cf882bbfa43f5ba499b63b6173674890ad00f9041d46f","abstract_canon_sha256":"196b3c99577fc34f38073178905f660f06d78a55fc33225c29a7a30e9a917c95"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:12:01.141103Z","signature_b64":"3SZDgCmBE09aM6DS0VneaIjIrAPAdd0ZrA4DHbNISz+xkt7UXMg8EvvSCGLEjmS85RDN0RzzBJkfraGAqa8VDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"279c025619167fae5f35ba784c624f75bc6dc98fb28d0e5c48d824f0976f100b","last_reissued_at":"2026-07-05T07:12:01.140659Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:12:01.140659Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Proto-lm: A Prototypical Network-Based Framework for Built-in Interpretability in Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Saeed Hassanpour, Sean Xie, Soroush Vosoughi","submitted_at":"2023-11-03T05:55:32Z","abstract_excerpt":"Large Language Models (LLMs) have significantly advanced the field of Natural Language Processing (NLP), but their lack of interpretability has been a major concern. Current methods for interpreting LLMs are post hoc, applied after inference time, and have limitations such as their focus on low-level features and lack of explainability at higher level text units. In this work, we introduce proto-lm, a prototypical network-based white-box framework that allows LLMs to learn immediately interpretable embeddings during the fine-tuning stage while maintaining competitive performance. Our method's "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.01732","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/2311.01732/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":"2311.01732","created_at":"2026-07-05T07:12:01.140719+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.01732v2","created_at":"2026-07-05T07:12:01.140719+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.01732","created_at":"2026-07-05T07:12:01.140719+00:00"},{"alias_kind":"pith_short_12","alias_value":"E6OAEVQZCZ72","created_at":"2026-07-05T07:12:01.140719+00:00"},{"alias_kind":"pith_short_16","alias_value":"E6OAEVQZCZ724XZV","created_at":"2026-07-05T07:12:01.140719+00:00"},{"alias_kind":"pith_short_8","alias_value":"E6OAEVQZ","created_at":"2026-07-05T07:12:01.140719+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.12169","citing_title":"Regulation of Language Models With Interpretability Will Likely Result In A Performance Trade-Off","ref_index":34,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/E6OAEVQZCZ724XZVXJ4EYYSPOW","json":"https://pith.science/pith/E6OAEVQZCZ724XZVXJ4EYYSPOW.json","graph_json":"https://pith.science/api/pith-number/E6OAEVQZCZ724XZVXJ4EYYSPOW/graph.json","events_json":"https://pith.science/api/pith-number/E6OAEVQZCZ724XZVXJ4EYYSPOW/events.json","paper":"https://pith.science/paper/E6OAEVQZ"},"agent_actions":{"view_html":"https://pith.science/pith/E6OAEVQZCZ724XZVXJ4EYYSPOW","download_json":"https://pith.science/pith/E6OAEVQZCZ724XZVXJ4EYYSPOW.json","view_paper":"https://pith.science/paper/E6OAEVQZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.01732&json=true","fetch_graph":"https://pith.science/api/pith-number/E6OAEVQZCZ724XZVXJ4EYYSPOW/graph.json","fetch_events":"https://pith.science/api/pith-number/E6OAEVQZCZ724XZVXJ4EYYSPOW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/E6OAEVQZCZ724XZVXJ4EYYSPOW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/E6OAEVQZCZ724XZVXJ4EYYSPOW/action/storage_attestation","attest_author":"https://pith.science/pith/E6OAEVQZCZ724XZVXJ4EYYSPOW/action/author_attestation","sign_citation":"https://pith.science/pith/E6OAEVQZCZ724XZVXJ4EYYSPOW/action/citation_signature","submit_replication":"https://pith.science/pith/E6OAEVQZCZ724XZVXJ4EYYSPOW/action/replication_record"}},"created_at":"2026-07-05T07:12:01.140719+00:00","updated_at":"2026-07-05T07:12:01.140719+00:00"}