{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:BBZCER552AZJLAAOZQZP6NQXIG","short_pith_number":"pith:BBZCER55","schema_version":"1.0","canonical_sha256":"08722247bdd03295800ecc32ff361741a2a4785077676e7707a02cf81b5cd479","source":{"kind":"arxiv","id":"2608.05872","version":1},"attestation_state":"computed","paper":{"title":"MACRO: Markov Chain Routing of Transformer Layers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Abtin Pourhadi, Akylgali Aitaza, Paul Swoboda, Pawe{\\l} Batorski, Przemys{\\l}aw Spurek","submitted_at":"2026-08-06T10:51:34Z","abstract_excerpt":"Standard Large Language Models (LLMs) execute layers sequentially. Dynamic layer routing, i.e. search for a different execution path through layers involving layer repetitions, skips and other moves, can improve performance. Existing routing approaches often require updating model weights, running expensive search loops per test instance, or demand ground-truth labels during inference. In this work, we propose Markov Chain Routing of Transformer Layers (MACRO), a framework that learns task-specific routes over LLM architectures without modifying underlying parameters. MACRO models layer routin"},"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":"2608.05872","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-08-06T10:51:34Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"a7a9fd92128824320ef8c5eff311aea08ed0fc5af1fefc66f688fb1a2930b447","abstract_canon_sha256":"40704761ff3a915517ceb7da8034aa568aa1d8ea177bfd1c022b6739bc8eae7d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-07T00:53:10.849358Z","signature_b64":"I2eZ1VARqZ+46uoBshBUA11eD4R7a0skoBo+sK6Rd2X3SBzXFKJuItvKjCrT5iqKXBncmu7z6M+nw6PlBb8CBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"08722247bdd03295800ecc32ff361741a2a4785077676e7707a02cf81b5cd479","last_reissued_at":"2026-08-07T00:53:10.847869Z","signature_status":"signed_v1","first_computed_at":"2026-08-07T00:53:10.847869Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MACRO: Markov Chain Routing of Transformer Layers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Abtin Pourhadi, Akylgali Aitaza, Paul Swoboda, Pawe{\\l} Batorski, Przemys{\\l}aw Spurek","submitted_at":"2026-08-06T10:51:34Z","abstract_excerpt":"Standard Large Language Models (LLMs) execute layers sequentially. Dynamic layer routing, i.e. search for a different execution path through layers involving layer repetitions, skips and other moves, can improve performance. Existing routing approaches often require updating model weights, running expensive search loops per test instance, or demand ground-truth labels during inference. In this work, we propose Markov Chain Routing of Transformer Layers (MACRO), a framework that learns task-specific routes over LLM architectures without modifying underlying parameters. MACRO models layer routin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.05872","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/2608.05872/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":"2608.05872","created_at":"2026-08-07T00:53:10.849375+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.05872v1","created_at":"2026-08-07T00:53:10.849375+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.05872","created_at":"2026-08-07T00:53:10.849375+00:00"},{"alias_kind":"pith_short_12","alias_value":"BBZCER552AZJ","created_at":"2026-08-07T00:53:10.849375+00:00"},{"alias_kind":"pith_short_16","alias_value":"BBZCER552AZJLAAO","created_at":"2026-08-07T00:53:10.849375+00:00"},{"alias_kind":"pith_short_8","alias_value":"BBZCER55","created_at":"2026-08-07T00:53:10.849375+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/BBZCER552AZJLAAOZQZP6NQXIG","json":"https://pith.science/pith/BBZCER552AZJLAAOZQZP6NQXIG.json","graph_json":"https://pith.science/api/pith-number/BBZCER552AZJLAAOZQZP6NQXIG/graph.json","events_json":"https://pith.science/api/pith-number/BBZCER552AZJLAAOZQZP6NQXIG/events.json","paper":"https://pith.science/paper/BBZCER55"},"agent_actions":{"view_html":"https://pith.science/pith/BBZCER552AZJLAAOZQZP6NQXIG","download_json":"https://pith.science/pith/BBZCER552AZJLAAOZQZP6NQXIG.json","view_paper":"https://pith.science/paper/BBZCER55","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.05872&json=true","fetch_graph":"https://pith.science/api/pith-number/BBZCER552AZJLAAOZQZP6NQXIG/graph.json","fetch_events":"https://pith.science/api/pith-number/BBZCER552AZJLAAOZQZP6NQXIG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BBZCER552AZJLAAOZQZP6NQXIG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BBZCER552AZJLAAOZQZP6NQXIG/action/storage_attestation","attest_author":"https://pith.science/pith/BBZCER552AZJLAAOZQZP6NQXIG/action/author_attestation","sign_citation":"https://pith.science/pith/BBZCER552AZJLAAOZQZP6NQXIG/action/citation_signature","submit_replication":"https://pith.science/pith/BBZCER552AZJLAAOZQZP6NQXIG/action/replication_record"}},"created_at":"2026-08-07T00:53:10.849375+00:00","updated_at":"2026-08-07T00:53:10.849375+00:00"}