{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:SWH5EN7OZBYXK4TNKZKD3QKIOG","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"abbf9c7f68ef32a59dac45c943b0cb349704056d539fc128c8c19f3b2cff57fb","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-11T21:41:21Z","title_canon_sha256":"9f0bb0f883fba2f52f3c58659024484e6ddf4892cc29bfe032f93e67d1ef5ba6"},"schema_version":"1.0","source":{"id":"2605.11262","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.11262","created_at":"2026-05-20T02:05:45Z"},{"alias_kind":"arxiv_version","alias_value":"2605.11262v2","created_at":"2026-05-20T02:05:45Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.11262","created_at":"2026-05-20T02:05:45Z"},{"alias_kind":"pith_short_12","alias_value":"SWH5EN7OZBYX","created_at":"2026-05-20T02:05:45Z"},{"alias_kind":"pith_short_16","alias_value":"SWH5EN7OZBYXK4TN","created_at":"2026-05-20T02:05:45Z"},{"alias_kind":"pith_short_8","alias_value":"SWH5EN7O","created_at":"2026-05-20T02:05:45Z"}],"graph_snapshots":[{"event_id":"sha256:adcaf8561491fffb9b500dd99b8d1f77f9841cc2dfebd4d94886f7293f5be7c0","target":"graph","created_at":"2026-05-20T02:05:45Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Across 36 datasets in time-series forecasting and tabular prediction, latent chain-of-thought improves over the baseline on 8/9 time-series datasets (+10.99% average gain) and 22/27 tabular datasets (+5.31% average gain). Across both settings, the CoT models perform the best on average."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the observed gains are attributable to the latent chain-of-thought mechanism rather than incidental effects of the recurrent architecture or token compression, and that the gains generalize beyond the 36 chosen datasets and the specific transformer backbone used."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Latent chain-of-thought via recurrent feedback tokens improves average performance of structured-data transformers on time-series forecasting and tabular prediction."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Latent chain-of-thought via recurrent feedback tokens improves transformer performance on time-series forecasting and tabular prediction."}],"snapshot_sha256":"809c745b81c3c31b80020bbcc52821cb3720a43e7dd175c6434d061bfa65f733"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"88be73c0dcace7f2a750dc157e773d0d2151fb00d8f55b2ff7e35f3fe33fa191"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-19T12:39:32.204627Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_title_agreement","ran_at":"2026-05-19T10:01:17.285704Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T08:37:05.507499Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.11262/integrity.json","findings":[],"snapshot_sha256":"8aff626376bba00443f92e7cf6ce5eb43e0aaa6b1fb6e9a2cfd6dae3aacfecd0","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Chain-of-thought and more broadly test-time compute are known to augment the expressive capabilities of language models and have led to major innovations in reasoning. Motivated by this success, this paper explores latent chain-of-thought as well as the impact of depth and looping for time-series and tabular data. We propose a recurrent scheme in which a structured-data transformer, after an initial forward pass, compresses its query-position hidden states into feedback tokens that are appended to the input and processed again, allowing multiple rounds of latent computation before prediction. ","authors_text":"Carson Dudley, Samet Oymak","cross_cats":[],"headline":"Latent chain-of-thought via recurrent feedback tokens improves transformer performance on time-series forecasting and tabular prediction.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-11T21:41:21Z","title":"Latent Chain-of-Thought Improves Structured-Data Transformers"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.11262","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-13T01:44:46.292821Z","id":"33656d52-67b7-43ea-ab99-bc13e1562b6d","model_set":{"reader":"grok-4.3"},"one_line_summary":"Latent chain-of-thought via recurrent feedback tokens improves average performance of structured-data transformers on time-series forecasting and tabular prediction.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Latent chain-of-thought via recurrent feedback tokens improves transformer performance on time-series forecasting and tabular prediction.","strongest_claim":"Across 36 datasets in time-series forecasting and tabular prediction, latent chain-of-thought improves over the baseline on 8/9 time-series datasets (+10.99% average gain) and 22/27 tabular datasets (+5.31% average gain). Across both settings, the CoT models perform the best on average.","weakest_assumption":"That the observed gains are attributable to the latent chain-of-thought mechanism rather than incidental effects of the recurrent architecture or token compression, and that the gains generalize beyond the 36 chosen datasets and the specific transformer backbone used."}},"verdict_id":"33656d52-67b7-43ea-ab99-bc13e1562b6d"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c9f699ba3b7e7842283774c1859bd7703204f7a568f446d2ba314f876c4481ff","target":"record","created_at":"2026-05-20T02:05:45Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"abbf9c7f68ef32a59dac45c943b0cb349704056d539fc128c8c19f3b2cff57fb","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-11T21:41:21Z","title_canon_sha256":"9f0bb0f883fba2f52f3c58659024484e6ddf4892cc29bfe032f93e67d1ef5ba6"},"schema_version":"1.0","source":{"id":"2605.11262","kind":"arxiv","version":2}},"canonical_sha256":"958fd237eec87175726d56543dc148719603bce3086a119ea7e75f444ff2ed75","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"958fd237eec87175726d56543dc148719603bce3086a119ea7e75f444ff2ed75","first_computed_at":"2026-05-20T02:05:45.888846Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T02:05:45.888846Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"3HEI0++2/pLWo0FN3SBNtIxP8A1WLLH/Z/QdG7UPM0l6t0Sy+RUkxnMlqPq961QoK/DLRIkEmKXfImDqSgaXAQ==","signature_status":"signed_v1","signed_at":"2026-05-20T02:05:45.889648Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.11262","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c9f699ba3b7e7842283774c1859bd7703204f7a568f446d2ba314f876c4481ff","sha256:adcaf8561491fffb9b500dd99b8d1f77f9841cc2dfebd4d94886f7293f5be7c0"],"state_sha256":"d1541d3b8b0e15747b87caaddc092532140734730dce38d4e50b3a28e59136c5"}