{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:5E6XFIIJCOX6L6PGS4XMP7UWSZ","short_pith_number":"pith:5E6XFIIJ","schema_version":"1.0","canonical_sha256":"e93d72a10913afe5f9e6972ec7fe9696592c603cf1c83e66a156fac870696ff3","source":{"kind":"arxiv","id":"2601.01754","version":3},"attestation_state":"computed","paper":{"title":"Context-Free Recognition with Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CC","cs.CL","cs.FL"],"primary_cat":"cs.LG","authors_text":"Anej Svete, Ryan Cotterell, Selim Jerad, Sophie Hao, William Merrill","submitted_at":"2026-01-05T03:14:23Z","abstract_excerpt":"Transformers excel empirically on tasks that process well-formed inputs according to some grammar, such as natural language and code. However, it remains unclear how they can process grammatical syntax. In fact, under standard complexity conjectures, standard transformers cannot recognize context-free languages (CFLs), a canonical formalism to describe syntax, or even regular languages, a subclass of CFLs. Past work has shown that $\\mathcal{O}(\\log(N))$ looping layers (w.r.t. input length $N$) allow transformers to recognize regular languages, but the question of context-free recognition with "},"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":"2601.01754","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-01-05T03:14:23Z","cross_cats_sorted":["cs.CC","cs.CL","cs.FL"],"title_canon_sha256":"53dab27aef552dc96145e98bc4647f51f6ce99718bf58523b0dc105c4c461297","abstract_canon_sha256":"1484aae81a61df2aa13a6caf1b585de6e327289917a5c76084f9298021e7babe"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-01T01:02:28.660755Z","signature_b64":"cC+lokIHe7lQwCp9O9Wzd6RUkWlHNo+H0LOhpmEEEh6N/kSBg4uirACX9msri4e2RYukJ0jrjYLtbn0fEGUAAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e93d72a10913afe5f9e6972ec7fe9696592c603cf1c83e66a156fac870696ff3","last_reissued_at":"2026-06-01T01:02:28.659679Z","signature_status":"signed_v1","first_computed_at":"2026-06-01T01:02:28.659679Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Context-Free Recognition with Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CC","cs.CL","cs.FL"],"primary_cat":"cs.LG","authors_text":"Anej Svete, Ryan Cotterell, Selim Jerad, Sophie Hao, William Merrill","submitted_at":"2026-01-05T03:14:23Z","abstract_excerpt":"Transformers excel empirically on tasks that process well-formed inputs according to some grammar, such as natural language and code. However, it remains unclear how they can process grammatical syntax. In fact, under standard complexity conjectures, standard transformers cannot recognize context-free languages (CFLs), a canonical formalism to describe syntax, or even regular languages, a subclass of CFLs. Past work has shown that $\\mathcal{O}(\\log(N))$ looping layers (w.r.t. input length $N$) allow transformers to recognize regular languages, but the question of context-free recognition with "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.01754","kind":"arxiv","version":3},"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/2601.01754/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":"2601.01754","created_at":"2026-06-01T01:02:28.659838+00:00"},{"alias_kind":"arxiv_version","alias_value":"2601.01754v3","created_at":"2026-06-01T01:02:28.659838+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.01754","created_at":"2026-06-01T01:02:28.659838+00:00"},{"alias_kind":"pith_short_12","alias_value":"5E6XFIIJCOX6","created_at":"2026-06-01T01:02:28.659838+00:00"},{"alias_kind":"pith_short_16","alias_value":"5E6XFIIJCOX6L6PG","created_at":"2026-06-01T01:02:28.659838+00:00"},{"alias_kind":"pith_short_8","alias_value":"5E6XFIIJ","created_at":"2026-06-01T01:02:28.659838+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2606.31779","citing_title":"Bridging the Gap Between Latent and Explicit Reasoning with Looped Transformers","ref_index":140,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5E6XFIIJCOX6L6PGS4XMP7UWSZ","json":"https://pith.science/pith/5E6XFIIJCOX6L6PGS4XMP7UWSZ.json","graph_json":"https://pith.science/api/pith-number/5E6XFIIJCOX6L6PGS4XMP7UWSZ/graph.json","events_json":"https://pith.science/api/pith-number/5E6XFIIJCOX6L6PGS4XMP7UWSZ/events.json","paper":"https://pith.science/paper/5E6XFIIJ"},"agent_actions":{"view_html":"https://pith.science/pith/5E6XFIIJCOX6L6PGS4XMP7UWSZ","download_json":"https://pith.science/pith/5E6XFIIJCOX6L6PGS4XMP7UWSZ.json","view_paper":"https://pith.science/paper/5E6XFIIJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2601.01754&json=true","fetch_graph":"https://pith.science/api/pith-number/5E6XFIIJCOX6L6PGS4XMP7UWSZ/graph.json","fetch_events":"https://pith.science/api/pith-number/5E6XFIIJCOX6L6PGS4XMP7UWSZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5E6XFIIJCOX6L6PGS4XMP7UWSZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5E6XFIIJCOX6L6PGS4XMP7UWSZ/action/storage_attestation","attest_author":"https://pith.science/pith/5E6XFIIJCOX6L6PGS4XMP7UWSZ/action/author_attestation","sign_citation":"https://pith.science/pith/5E6XFIIJCOX6L6PGS4XMP7UWSZ/action/citation_signature","submit_replication":"https://pith.science/pith/5E6XFIIJCOX6L6PGS4XMP7UWSZ/action/replication_record"}},"created_at":"2026-06-01T01:02:28.659838+00:00","updated_at":"2026-06-01T01:02:28.659838+00:00"}