{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:ULQIJHMTA63IZOBIMA3DJ62XWT","short_pith_number":"pith:ULQIJHMT","schema_version":"1.0","canonical_sha256":"a2e0849d9307b68cb828603634fb57b4fe09cb2027857edcf02194679e70a064","source":{"kind":"arxiv","id":"2203.00633","version":2},"attestation_state":"computed","paper":{"title":"Transformer Grammars: Augmenting Transformer Language Models with Syntactic Inductive Biases at Scale","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Adhiguna Kuncoro, Chris Dyer, Laurent Sartran, Milo\\v{s} Stanojevi\\'c, Phil Blunsom, Samuel Barrett","submitted_at":"2022-03-01T17:22:31Z","abstract_excerpt":"We introduce Transformer Grammars (TGs), a novel class of Transformer language models that combine (i) the expressive power, scalability, and strong performance of Transformers and (ii) recursive syntactic compositions, which here are implemented through a special attention mask and deterministic transformation of the linearized tree. We find that TGs outperform various strong baselines on sentence-level language modeling perplexity, as well as on multiple syntax-sensitive language modeling evaluation metrics. Additionally, we find that the recursive syntactic composition bottleneck which repr"},"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":"2203.00633","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-03-01T17:22:31Z","cross_cats_sorted":[],"title_canon_sha256":"b79eacaf7421e0a6b117265138609826660d6a566a25859d67a23bd6ba51ad73","abstract_canon_sha256":"530eb8b5a73a0f252870383ab1de0369de812b8a0d9bde17128c989bcc0e9345"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:22:44.136049Z","signature_b64":"9wekvqBEWz4nhv2hymW28HR3Hx1ic+OWOOL9kkz6G+8vljoZKS4cuBbjzVNt0FQmSQ+TqUSMw27/mZY5VldtAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a2e0849d9307b68cb828603634fb57b4fe09cb2027857edcf02194679e70a064","last_reissued_at":"2026-07-05T05:22:44.135509Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:22:44.135509Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Transformer Grammars: Augmenting Transformer Language Models with Syntactic Inductive Biases at Scale","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Adhiguna Kuncoro, Chris Dyer, Laurent Sartran, Milo\\v{s} Stanojevi\\'c, Phil Blunsom, Samuel Barrett","submitted_at":"2022-03-01T17:22:31Z","abstract_excerpt":"We introduce Transformer Grammars (TGs), a novel class of Transformer language models that combine (i) the expressive power, scalability, and strong performance of Transformers and (ii) recursive syntactic compositions, which here are implemented through a special attention mask and deterministic transformation of the linearized tree. We find that TGs outperform various strong baselines on sentence-level language modeling perplexity, as well as on multiple syntax-sensitive language modeling evaluation metrics. Additionally, we find that the recursive syntactic composition bottleneck which repr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.00633","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/2203.00633/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":"2203.00633","created_at":"2026-07-05T05:22:44.135568+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.00633v2","created_at":"2026-07-05T05:22:44.135568+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.00633","created_at":"2026-07-05T05:22:44.135568+00:00"},{"alias_kind":"pith_short_12","alias_value":"ULQIJHMTA63I","created_at":"2026-07-05T05:22:44.135568+00:00"},{"alias_kind":"pith_short_16","alias_value":"ULQIJHMTA63IZOBI","created_at":"2026-07-05T05:22:44.135568+00:00"},{"alias_kind":"pith_short_8","alias_value":"ULQIJHMT","created_at":"2026-07-05T05:22:44.135568+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2112.07874","citing_title":"Linguistic Frameworks Go Toe-to-Toe at Neuro-Symbolic Language Modeling","ref_index":51,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ULQIJHMTA63IZOBIMA3DJ62XWT","json":"https://pith.science/pith/ULQIJHMTA63IZOBIMA3DJ62XWT.json","graph_json":"https://pith.science/api/pith-number/ULQIJHMTA63IZOBIMA3DJ62XWT/graph.json","events_json":"https://pith.science/api/pith-number/ULQIJHMTA63IZOBIMA3DJ62XWT/events.json","paper":"https://pith.science/paper/ULQIJHMT"},"agent_actions":{"view_html":"https://pith.science/pith/ULQIJHMTA63IZOBIMA3DJ62XWT","download_json":"https://pith.science/pith/ULQIJHMTA63IZOBIMA3DJ62XWT.json","view_paper":"https://pith.science/paper/ULQIJHMT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.00633&json=true","fetch_graph":"https://pith.science/api/pith-number/ULQIJHMTA63IZOBIMA3DJ62XWT/graph.json","fetch_events":"https://pith.science/api/pith-number/ULQIJHMTA63IZOBIMA3DJ62XWT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ULQIJHMTA63IZOBIMA3DJ62XWT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ULQIJHMTA63IZOBIMA3DJ62XWT/action/storage_attestation","attest_author":"https://pith.science/pith/ULQIJHMTA63IZOBIMA3DJ62XWT/action/author_attestation","sign_citation":"https://pith.science/pith/ULQIJHMTA63IZOBIMA3DJ62XWT/action/citation_signature","submit_replication":"https://pith.science/pith/ULQIJHMTA63IZOBIMA3DJ62XWT/action/replication_record"}},"created_at":"2026-07-05T05:22:44.135568+00:00","updated_at":"2026-07-05T05:22:44.135568+00:00"}