{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:GXIXYIDIBCYBZ24KRTIREDDPJQ","short_pith_number":"pith:GXIXYIDI","schema_version":"1.0","canonical_sha256":"35d17c206808b01ceb8a8cd1120c6f4c3209a7e1943c15a051ed21ca5fc87262","source":{"kind":"arxiv","id":"2310.10262","version":1},"attestation_state":"computed","paper":{"title":"Enhancing Interpretability using Human Similarity Judgements to Prune Word Embeddings","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aurelie Herbelot, Natalia Flechas Manrique, Uri Hasson, Wanqian Bao","submitted_at":"2023-10-16T10:38:49Z","abstract_excerpt":"Interpretability methods in NLP aim to provide insights into the semantics underlying specific system architectures. Focusing on word embeddings, we present a supervised-learning method that, for a given domain (e.g., sports, professions), identifies a subset of model features that strongly improve prediction of human similarity judgments. We show this method keeps only 20-40% of the original embeddings, for 8 independent semantic domains, and that it retains different feature sets across domains. We then present two approaches for interpreting the semantics of the retained features. The first"},"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":"2310.10262","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-10-16T10:38:49Z","cross_cats_sorted":[],"title_canon_sha256":"3ba803c326a72882ea836fa1fa2a5b24126a502cdc2013b15aff0476909ed850","abstract_canon_sha256":"e37354d2350fae0e6925f6db4ab78b2c55d40d6a9ff8bc7745889e11a401ba92"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:01:17.805083Z","signature_b64":"EV2j9zEMIaFqE6b1GduitRpkxJ+hRxFQFHihkZ9qv2ewhhJt6THVUtt8pAVn7l/VaHe87vIxbL3Gn8OQsv82DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"35d17c206808b01ceb8a8cd1120c6f4c3209a7e1943c15a051ed21ca5fc87262","last_reissued_at":"2026-07-05T07:01:17.804647Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:01:17.804647Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Enhancing Interpretability using Human Similarity Judgements to Prune Word Embeddings","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Aurelie Herbelot, Natalia Flechas Manrique, Uri Hasson, Wanqian Bao","submitted_at":"2023-10-16T10:38:49Z","abstract_excerpt":"Interpretability methods in NLP aim to provide insights into the semantics underlying specific system architectures. Focusing on word embeddings, we present a supervised-learning method that, for a given domain (e.g., sports, professions), identifies a subset of model features that strongly improve prediction of human similarity judgments. We show this method keeps only 20-40% of the original embeddings, for 8 independent semantic domains, and that it retains different feature sets across domains. We then present two approaches for interpreting the semantics of the retained features. The first"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.10262","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/2310.10262/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":"2310.10262","created_at":"2026-07-05T07:01:17.804699+00:00"},{"alias_kind":"arxiv_version","alias_value":"2310.10262v1","created_at":"2026-07-05T07:01:17.804699+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.10262","created_at":"2026-07-05T07:01:17.804699+00:00"},{"alias_kind":"pith_short_12","alias_value":"GXIXYIDIBCYB","created_at":"2026-07-05T07:01:17.804699+00:00"},{"alias_kind":"pith_short_16","alias_value":"GXIXYIDIBCYBZ24K","created_at":"2026-07-05T07:01:17.804699+00:00"},{"alias_kind":"pith_short_8","alias_value":"GXIXYIDI","created_at":"2026-07-05T07:01:17.804699+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2504.21047","citing_title":"Model Connectomes: A Generational Approach to Data-Efficient Language Models","ref_index":32,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GXIXYIDIBCYBZ24KRTIREDDPJQ","json":"https://pith.science/pith/GXIXYIDIBCYBZ24KRTIREDDPJQ.json","graph_json":"https://pith.science/api/pith-number/GXIXYIDIBCYBZ24KRTIREDDPJQ/graph.json","events_json":"https://pith.science/api/pith-number/GXIXYIDIBCYBZ24KRTIREDDPJQ/events.json","paper":"https://pith.science/paper/GXIXYIDI"},"agent_actions":{"view_html":"https://pith.science/pith/GXIXYIDIBCYBZ24KRTIREDDPJQ","download_json":"https://pith.science/pith/GXIXYIDIBCYBZ24KRTIREDDPJQ.json","view_paper":"https://pith.science/paper/GXIXYIDI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2310.10262&json=true","fetch_graph":"https://pith.science/api/pith-number/GXIXYIDIBCYBZ24KRTIREDDPJQ/graph.json","fetch_events":"https://pith.science/api/pith-number/GXIXYIDIBCYBZ24KRTIREDDPJQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GXIXYIDIBCYBZ24KRTIREDDPJQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GXIXYIDIBCYBZ24KRTIREDDPJQ/action/storage_attestation","attest_author":"https://pith.science/pith/GXIXYIDIBCYBZ24KRTIREDDPJQ/action/author_attestation","sign_citation":"https://pith.science/pith/GXIXYIDIBCYBZ24KRTIREDDPJQ/action/citation_signature","submit_replication":"https://pith.science/pith/GXIXYIDIBCYBZ24KRTIREDDPJQ/action/replication_record"}},"created_at":"2026-07-05T07:01:17.804699+00:00","updated_at":"2026-07-05T07:01:17.804699+00:00"}