{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:O26SUCG7EQHMDDU2ATOXMDIZHX","short_pith_number":"pith:O26SUCG7","schema_version":"1.0","canonical_sha256":"76bd2a08df240ec18e9a04dd760d193dda8a6e312b064d71d74c65a21334c30d","source":{"kind":"arxiv","id":"2605.28283","version":1},"attestation_state":"computed","paper":{"title":"PrunePath: Towards Highly Structured Sparse Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Yancheng Yuan, Zhexuan Gu, Zixun Fu","submitted_at":"2026-05-27T10:29:52Z","abstract_excerpt":"Feed-forward networks (FFNs) dominate the parameter count and computation of modern language models, yet existing pruning methods often struggle to convert sparsity into hardware-friendly inference efficiency gains. We introduce \\textbf{PrunePath}, a budget-adaptive structured sparsification framework for FFN layers. Built on MoEfication, PrunePath replaces independent expert-wise thresholding with a softmax-normalized routing distribution and activates important experts under a cumulative-mass threshold. This formulation imposes a token-level probability budget, enabling adaptive expert count"},"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":"2605.28283","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-27T10:29:52Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"c9b5fb7285caf6970d31b41229e6d39d5f9db092d2a918d3ad702d8f4211906b","abstract_canon_sha256":"d533dee13b1be5eadb69db98c5b44132a0ed88998a4756cf9e7bb48178a7e12a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:05:05.142138Z","signature_b64":"E3rHAaNqf9alV0K1bQUnKytfhBJQptuJk08Mj3k+XPKgPT49fG1VRMlHZFF3vmJhA5EMm0KlACviKGBx5rzQDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"76bd2a08df240ec18e9a04dd760d193dda8a6e312b064d71d74c65a21334c30d","last_reissued_at":"2026-05-28T01:05:05.141557Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:05:05.141557Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"PrunePath: Towards Highly Structured Sparse Language Models","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Yancheng Yuan, Zhexuan Gu, Zixun Fu","submitted_at":"2026-05-27T10:29:52Z","abstract_excerpt":"Feed-forward networks (FFNs) dominate the parameter count and computation of modern language models, yet existing pruning methods often struggle to convert sparsity into hardware-friendly inference efficiency gains. We introduce \\textbf{PrunePath}, a budget-adaptive structured sparsification framework for FFN layers. Built on MoEfication, PrunePath replaces independent expert-wise thresholding with a softmax-normalized routing distribution and activates important experts under a cumulative-mass threshold. This formulation imposes a token-level probability budget, enabling adaptive expert count"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28283","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/2605.28283/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":"2605.28283","created_at":"2026-05-28T01:05:05.141629+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.28283v1","created_at":"2026-05-28T01:05:05.141629+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28283","created_at":"2026-05-28T01:05:05.141629+00:00"},{"alias_kind":"pith_short_12","alias_value":"O26SUCG7EQHM","created_at":"2026-05-28T01:05:05.141629+00:00"},{"alias_kind":"pith_short_16","alias_value":"O26SUCG7EQHMDDU2","created_at":"2026-05-28T01:05:05.141629+00:00"},{"alias_kind":"pith_short_8","alias_value":"O26SUCG7","created_at":"2026-05-28T01:05:05.141629+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/O26SUCG7EQHMDDU2ATOXMDIZHX","json":"https://pith.science/pith/O26SUCG7EQHMDDU2ATOXMDIZHX.json","graph_json":"https://pith.science/api/pith-number/O26SUCG7EQHMDDU2ATOXMDIZHX/graph.json","events_json":"https://pith.science/api/pith-number/O26SUCG7EQHMDDU2ATOXMDIZHX/events.json","paper":"https://pith.science/paper/O26SUCG7"},"agent_actions":{"view_html":"https://pith.science/pith/O26SUCG7EQHMDDU2ATOXMDIZHX","download_json":"https://pith.science/pith/O26SUCG7EQHMDDU2ATOXMDIZHX.json","view_paper":"https://pith.science/paper/O26SUCG7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.28283&json=true","fetch_graph":"https://pith.science/api/pith-number/O26SUCG7EQHMDDU2ATOXMDIZHX/graph.json","fetch_events":"https://pith.science/api/pith-number/O26SUCG7EQHMDDU2ATOXMDIZHX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/O26SUCG7EQHMDDU2ATOXMDIZHX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/O26SUCG7EQHMDDU2ATOXMDIZHX/action/storage_attestation","attest_author":"https://pith.science/pith/O26SUCG7EQHMDDU2ATOXMDIZHX/action/author_attestation","sign_citation":"https://pith.science/pith/O26SUCG7EQHMDDU2ATOXMDIZHX/action/citation_signature","submit_replication":"https://pith.science/pith/O26SUCG7EQHMDDU2ATOXMDIZHX/action/replication_record"}},"created_at":"2026-05-28T01:05:05.141629+00:00","updated_at":"2026-05-28T01:05:05.141629+00:00"}