{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:D3UOW6Q2MSG7YUXT767EQC2QPO","short_pith_number":"pith:D3UOW6Q2","schema_version":"1.0","canonical_sha256":"1ee8eb7a1a648dfc52f3ffbe480b507bb44101db0b1ee8dec96e353e4c0cbf20","source":{"kind":"arxiv","id":"2503.06291","version":1},"attestation_state":"computed","paper":{"title":"IteRABRe: Iterative Recovery-Aided Block Reduction","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alham Fikri Aji, Haiyue Song, Haryo Akbarianto Wibowo, Hideki Tanaka, Masao Utiyama, Raj Dabre","submitted_at":"2025-03-08T17:46:01Z","abstract_excerpt":"Large Language Models (LLMs) have grown increasingly expensive to deploy, driving the need for effective model compression techniques. While block pruning offers a straightforward approach to reducing model size, existing methods often struggle to maintain performance or require substantial computational resources for recovery. We present IteRABRe, a simple yet effective iterative pruning method that achieves superior compression results while requiring minimal computational resources. Using only 2.5M tokens for recovery, our method outperforms baseline approaches by ~3% on average when compre"},"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":"2503.06291","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-08T17:46:01Z","cross_cats_sorted":[],"title_canon_sha256":"4cc38d38ec2d8a5ea350f87de9ec9b1644cca404391608989ae91d2ff76f268a","abstract_canon_sha256":"b2212472b75d8f0c9783ee2067ed3a29489887c1a74170e7553fb37b4f446cd3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:27:29.946517Z","signature_b64":"zcKXpwHSUZf7YMopnyKvr2ggGJwqMIFBcK2ClugvWpuJ90NVW83Q3KSicg/vaPOWx2BNYsH1zQETW9eprOtiBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1ee8eb7a1a648dfc52f3ffbe480b507bb44101db0b1ee8dec96e353e4c0cbf20","last_reissued_at":"2026-07-05T10:27:29.946039Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:27:29.946039Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"IteRABRe: Iterative Recovery-Aided Block Reduction","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Alham Fikri Aji, Haiyue Song, Haryo Akbarianto Wibowo, Hideki Tanaka, Masao Utiyama, Raj Dabre","submitted_at":"2025-03-08T17:46:01Z","abstract_excerpt":"Large Language Models (LLMs) have grown increasingly expensive to deploy, driving the need for effective model compression techniques. While block pruning offers a straightforward approach to reducing model size, existing methods often struggle to maintain performance or require substantial computational resources for recovery. We present IteRABRe, a simple yet effective iterative pruning method that achieves superior compression results while requiring minimal computational resources. Using only 2.5M tokens for recovery, our method outperforms baseline approaches by ~3% on average when compre"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.06291","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/2503.06291/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":"2503.06291","created_at":"2026-07-05T10:27:29.946095+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.06291v1","created_at":"2026-07-05T10:27:29.946095+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.06291","created_at":"2026-07-05T10:27:29.946095+00:00"},{"alias_kind":"pith_short_12","alias_value":"D3UOW6Q2MSG7","created_at":"2026-07-05T10:27:29.946095+00:00"},{"alias_kind":"pith_short_16","alias_value":"D3UOW6Q2MSG7YUXT","created_at":"2026-07-05T10:27:29.946095+00:00"},{"alias_kind":"pith_short_8","alias_value":"D3UOW6Q2","created_at":"2026-07-05T10:27:29.946095+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.20237","citing_title":"Efficient Speech Translation through Model Compression and Knowledge Distillation","ref_index":30,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/D3UOW6Q2MSG7YUXT767EQC2QPO","json":"https://pith.science/pith/D3UOW6Q2MSG7YUXT767EQC2QPO.json","graph_json":"https://pith.science/api/pith-number/D3UOW6Q2MSG7YUXT767EQC2QPO/graph.json","events_json":"https://pith.science/api/pith-number/D3UOW6Q2MSG7YUXT767EQC2QPO/events.json","paper":"https://pith.science/paper/D3UOW6Q2"},"agent_actions":{"view_html":"https://pith.science/pith/D3UOW6Q2MSG7YUXT767EQC2QPO","download_json":"https://pith.science/pith/D3UOW6Q2MSG7YUXT767EQC2QPO.json","view_paper":"https://pith.science/paper/D3UOW6Q2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.06291&json=true","fetch_graph":"https://pith.science/api/pith-number/D3UOW6Q2MSG7YUXT767EQC2QPO/graph.json","fetch_events":"https://pith.science/api/pith-number/D3UOW6Q2MSG7YUXT767EQC2QPO/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/D3UOW6Q2MSG7YUXT767EQC2QPO/action/timestamp_anchor","attest_storage":"https://pith.science/pith/D3UOW6Q2MSG7YUXT767EQC2QPO/action/storage_attestation","attest_author":"https://pith.science/pith/D3UOW6Q2MSG7YUXT767EQC2QPO/action/author_attestation","sign_citation":"https://pith.science/pith/D3UOW6Q2MSG7YUXT767EQC2QPO/action/citation_signature","submit_replication":"https://pith.science/pith/D3UOW6Q2MSG7YUXT767EQC2QPO/action/replication_record"}},"created_at":"2026-07-05T10:27:29.946095+00:00","updated_at":"2026-07-05T10:27:29.946095+00:00"}