{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NDJEEPKI2UMKEENQP4M4BQDZHZ","short_pith_number":"pith:NDJEEPKI","schema_version":"1.0","canonical_sha256":"68d2423d48d518a211b07f19c0c0793e7deecd3b550875f4e1a6a4e86e7819d9","source":{"kind":"arxiv","id":"2605.23872","version":1},"attestation_state":"computed","paper":{"title":"Training-Free Looped Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NA","math.NA","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chen Liang, Jonathan Li, Lizhang Chen, Ni Lao, Qiang Liu","submitted_at":"2026-05-22T17:31:16Z","abstract_excerpt":"We introduce training-free looped transformers, in which a lightweight inference-time wrapper loops a contiguous mid-stack block of layers of a frozen checkpoint without additional fine-tuning, continued training, or architectural changes. Unlike prior looped transformer methods that train with the looped structure end-to-end, we retrofit recurrence onto pretrained models at test time. We show that naive block reapplication usually degrades performance, highlighting the importance of the loop application strategy. Motivated by viewing a pre-norm transformer block as a forward Euler step on an "},"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":"2605.23872","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-22T17:31:16Z","cross_cats_sorted":["cs.NA","math.NA","stat.ML"],"title_canon_sha256":"e94a9ac1d2b5c7aba1c18e9d200c11bc69f29ab8f2b2af8a4cffbf9c4d19632e","abstract_canon_sha256":"cc3ea13304781cd71cc795fd62b65fff8085b11bb3f994bed66cc081c3a419ba"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-25T02:02:37.026770Z","signature_b64":"grHIp7b6WCyQ0MldYhnQhS7DDmo6qpoA5cmZaMtNar8W0Yz07g1YXe9umwzTpHvTrPl/tjCgD/YWmkdA9sUvAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"68d2423d48d518a211b07f19c0c0793e7deecd3b550875f4e1a6a4e86e7819d9","last_reissued_at":"2026-05-25T02:02:37.026199Z","signature_status":"signed_v1","first_computed_at":"2026-05-25T02:02:37.026199Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Training-Free Looped Transformers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.NA","math.NA","stat.ML"],"primary_cat":"cs.LG","authors_text":"Chen Liang, Jonathan Li, Lizhang Chen, Ni Lao, Qiang Liu","submitted_at":"2026-05-22T17:31:16Z","abstract_excerpt":"We introduce training-free looped transformers, in which a lightweight inference-time wrapper loops a contiguous mid-stack block of layers of a frozen checkpoint without additional fine-tuning, continued training, or architectural changes. Unlike prior looped transformer methods that train with the looped structure end-to-end, we retrofit recurrence onto pretrained models at test time. We show that naive block reapplication usually degrades performance, highlighting the importance of the loop application strategy. Motivated by viewing a pre-norm transformer block as a forward Euler step on an "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.23872","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/2605.23872/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":"2605.23872","created_at":"2026-05-25T02:02:37.026285+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.23872v1","created_at":"2026-05-25T02:02:37.026285+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.23872","created_at":"2026-05-25T02:02:37.026285+00:00"},{"alias_kind":"pith_short_12","alias_value":"NDJEEPKI2UMK","created_at":"2026-05-25T02:02:37.026285+00:00"},{"alias_kind":"pith_short_16","alias_value":"NDJEEPKI2UMKEENQ","created_at":"2026-05-25T02:02:37.026285+00:00"},{"alias_kind":"pith_short_8","alias_value":"NDJEEPKI","created_at":"2026-05-25T02:02:37.026285+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/NDJEEPKI2UMKEENQP4M4BQDZHZ","json":"https://pith.science/pith/NDJEEPKI2UMKEENQP4M4BQDZHZ.json","graph_json":"https://pith.science/api/pith-number/NDJEEPKI2UMKEENQP4M4BQDZHZ/graph.json","events_json":"https://pith.science/api/pith-number/NDJEEPKI2UMKEENQP4M4BQDZHZ/events.json","paper":"https://pith.science/paper/NDJEEPKI"},"agent_actions":{"view_html":"https://pith.science/pith/NDJEEPKI2UMKEENQP4M4BQDZHZ","download_json":"https://pith.science/pith/NDJEEPKI2UMKEENQP4M4BQDZHZ.json","view_paper":"https://pith.science/paper/NDJEEPKI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.23872&json=true","fetch_graph":"https://pith.science/api/pith-number/NDJEEPKI2UMKEENQP4M4BQDZHZ/graph.json","fetch_events":"https://pith.science/api/pith-number/NDJEEPKI2UMKEENQP4M4BQDZHZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NDJEEPKI2UMKEENQP4M4BQDZHZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NDJEEPKI2UMKEENQP4M4BQDZHZ/action/storage_attestation","attest_author":"https://pith.science/pith/NDJEEPKI2UMKEENQP4M4BQDZHZ/action/author_attestation","sign_citation":"https://pith.science/pith/NDJEEPKI2UMKEENQP4M4BQDZHZ/action/citation_signature","submit_replication":"https://pith.science/pith/NDJEEPKI2UMKEENQP4M4BQDZHZ/action/replication_record"}},"created_at":"2026-05-25T02:02:37.026285+00:00","updated_at":"2026-05-25T02:02:37.026285+00:00"}