{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:4ZOPKBUCC7BPAT4TUMUPWJZCCF","short_pith_number":"pith:4ZOPKBUC","schema_version":"1.0","canonical_sha256":"e65cf5068217c2f04f93a328fb2722114cd01171154462578a2c75b2b61999b7","source":{"kind":"arxiv","id":"2605.30233","version":1},"attestation_state":"computed","paper":{"title":"Do Language Models Track Entities Across State Changes?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aaron Mueller, Derry Wijaya, Gabriel Franco, Najoung Kim, Qiao Zhao, Sebastian Schuster, Zilu Tang","submitted_at":"2026-05-28T17:03:42Z","abstract_excerpt":"Entity tracking (ET), the ability to keep track of states, is a fundamental skill that underlies complex reasoning. An increasing amount of work investigates how transformer language models (LMs) solve entity binding $\\textit{without}$ state changes. However, there is limited understanding of how non-toy LMs address ET problems of realistic difficulties expressed in natural language. To this end, we investigate the mechanisms underlying ET in more complex scenarios featuring multiple state-changing operations. We find that LMs do not incrementally track world states across tokens or query-rele"},"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":"2605.30233","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-28T17:03:42Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b4353121b9c9f1268e7feb170b496207f724dc467d2f83cef7ccd6536843459b","abstract_canon_sha256":"b8b228e3ca97ddafbb7be48913b1287ab50727f17b6ff11aa02067578872984a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T02:06:13.535743Z","signature_b64":"Y1hvF3DTvOrkIWjKvCONZM1xviionsPaK5a6FVPC6Ob7r9/Enhysg44/QXVCyzQBeY+UeUI+OaCe5sPju8TtDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e65cf5068217c2f04f93a328fb2722114cd01171154462578a2c75b2b61999b7","last_reissued_at":"2026-05-29T02:06:13.535366Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T02:06:13.535366Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do Language Models Track Entities Across State Changes?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Aaron Mueller, Derry Wijaya, Gabriel Franco, Najoung Kim, Qiao Zhao, Sebastian Schuster, Zilu Tang","submitted_at":"2026-05-28T17:03:42Z","abstract_excerpt":"Entity tracking (ET), the ability to keep track of states, is a fundamental skill that underlies complex reasoning. An increasing amount of work investigates how transformer language models (LMs) solve entity binding $\\textit{without}$ state changes. However, there is limited understanding of how non-toy LMs address ET problems of realistic difficulties expressed in natural language. To this end, we investigate the mechanisms underlying ET in more complex scenarios featuring multiple state-changing operations. We find that LMs do not incrementally track world states across tokens or query-rele"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.30233","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/2605.30233/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":"2605.30233","created_at":"2026-05-29T02:06:13.535432+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.30233v1","created_at":"2026-05-29T02:06:13.535432+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.30233","created_at":"2026-05-29T02:06:13.535432+00:00"},{"alias_kind":"pith_short_12","alias_value":"4ZOPKBUCC7BP","created_at":"2026-05-29T02:06:13.535432+00:00"},{"alias_kind":"pith_short_16","alias_value":"4ZOPKBUCC7BPAT4T","created_at":"2026-05-29T02:06:13.535432+00:00"},{"alias_kind":"pith_short_8","alias_value":"4ZOPKBUC","created_at":"2026-05-29T02:06:13.535432+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4ZOPKBUCC7BPAT4TUMUPWJZCCF","json":"https://pith.science/pith/4ZOPKBUCC7BPAT4TUMUPWJZCCF.json","graph_json":"https://pith.science/api/pith-number/4ZOPKBUCC7BPAT4TUMUPWJZCCF/graph.json","events_json":"https://pith.science/api/pith-number/4ZOPKBUCC7BPAT4TUMUPWJZCCF/events.json","paper":"https://pith.science/paper/4ZOPKBUC"},"agent_actions":{"view_html":"https://pith.science/pith/4ZOPKBUCC7BPAT4TUMUPWJZCCF","download_json":"https://pith.science/pith/4ZOPKBUCC7BPAT4TUMUPWJZCCF.json","view_paper":"https://pith.science/paper/4ZOPKBUC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.30233&json=true","fetch_graph":"https://pith.science/api/pith-number/4ZOPKBUCC7BPAT4TUMUPWJZCCF/graph.json","fetch_events":"https://pith.science/api/pith-number/4ZOPKBUCC7BPAT4TUMUPWJZCCF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4ZOPKBUCC7BPAT4TUMUPWJZCCF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4ZOPKBUCC7BPAT4TUMUPWJZCCF/action/storage_attestation","attest_author":"https://pith.science/pith/4ZOPKBUCC7BPAT4TUMUPWJZCCF/action/author_attestation","sign_citation":"https://pith.science/pith/4ZOPKBUCC7BPAT4TUMUPWJZCCF/action/citation_signature","submit_replication":"https://pith.science/pith/4ZOPKBUCC7BPAT4TUMUPWJZCCF/action/replication_record"}},"created_at":"2026-05-29T02:06:13.535432+00:00","updated_at":"2026-05-29T02:06:13.535432+00:00"}