{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:RHVEXAKNPXGMCBRQSRY7RMLMEW","short_pith_number":"pith:RHVEXAKN","schema_version":"1.0","canonical_sha256":"89ea4b814d7dccc106309471f8b16c25b4b00bff4bec3436509c9fd75daa88b4","source":{"kind":"arxiv","id":"2503.07605","version":1},"attestation_state":"computed","paper":{"title":"SEAP: Training-free Sparse Expert Activation Pruning Unlock the Brainpower of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bo Tang, Feiyu Xiong, Hanyu Wang, Huayi Lai, Jiawei Yang, Jihao Zhao, Shichao Song, Simin Niu, Xun Liang, Zhiyu Li","submitted_at":"2025-03-10T17:59:03Z","abstract_excerpt":"Large Language Models have achieved remarkable success across various natural language processing tasks, yet their high computational cost during inference remains a major bottleneck. This paper introduces Sparse Expert Activation Pruning (SEAP), a training-free pruning method that selectively retains task-relevant parameters to reduce inference overhead. Inspired by the clustering patterns of hidden states and activations in LLMs, SEAP identifies task-specific expert activation patterns and prunes the model while preserving task performance and enhancing computational efficiency. Experimental"},"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":"2503.07605","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-03-10T17:59:03Z","cross_cats_sorted":[],"title_canon_sha256":"ef3b5203449b3851306dd6e909fd62cec2c0942a95546fb304e6af67318134b0","abstract_canon_sha256":"60802ea380be57ab16d5e420d0e41380b448dd119512d64419f5513132bfc0c6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:28:08.029198Z","signature_b64":"9TS2Bj4IdF8tOjI/X2PAQ9uauA1GCl94HqHh/Vsj5xcIBGGTN3GbuvFXIzrsz9pkXcfaGRID/OMKwEfFkiZiCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"89ea4b814d7dccc106309471f8b16c25b4b00bff4bec3436509c9fd75daa88b4","last_reissued_at":"2026-07-05T10:28:08.028468Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:28:08.028468Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SEAP: Training-free Sparse Expert Activation Pruning Unlock the Brainpower of Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bo Tang, Feiyu Xiong, Hanyu Wang, Huayi Lai, Jiawei Yang, Jihao Zhao, Shichao Song, Simin Niu, Xun Liang, Zhiyu Li","submitted_at":"2025-03-10T17:59:03Z","abstract_excerpt":"Large Language Models have achieved remarkable success across various natural language processing tasks, yet their high computational cost during inference remains a major bottleneck. This paper introduces Sparse Expert Activation Pruning (SEAP), a training-free pruning method that selectively retains task-relevant parameters to reduce inference overhead. Inspired by the clustering patterns of hidden states and activations in LLMs, SEAP identifies task-specific expert activation patterns and prunes the model while preserving task performance and enhancing computational efficiency. Experimental"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.07605","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/2503.07605/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":"2503.07605","created_at":"2026-07-05T10:28:08.028557+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.07605v1","created_at":"2026-07-05T10:28:08.028557+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.07605","created_at":"2026-07-05T10:28:08.028557+00:00"},{"alias_kind":"pith_short_12","alias_value":"RHVEXAKNPXGM","created_at":"2026-07-05T10:28:08.028557+00:00"},{"alias_kind":"pith_short_16","alias_value":"RHVEXAKNPXGMCBRQ","created_at":"2026-07-05T10:28:08.028557+00:00"},{"alias_kind":"pith_short_8","alias_value":"RHVEXAKN","created_at":"2026-07-05T10:28:08.028557+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.09886","citing_title":"SHAPE: Coalition-Aware Expert Pruning for Sparse Mixture-of-Experts LLMs","ref_index":14,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RHVEXAKNPXGMCBRQSRY7RMLMEW","json":"https://pith.science/pith/RHVEXAKNPXGMCBRQSRY7RMLMEW.json","graph_json":"https://pith.science/api/pith-number/RHVEXAKNPXGMCBRQSRY7RMLMEW/graph.json","events_json":"https://pith.science/api/pith-number/RHVEXAKNPXGMCBRQSRY7RMLMEW/events.json","paper":"https://pith.science/paper/RHVEXAKN"},"agent_actions":{"view_html":"https://pith.science/pith/RHVEXAKNPXGMCBRQSRY7RMLMEW","download_json":"https://pith.science/pith/RHVEXAKNPXGMCBRQSRY7RMLMEW.json","view_paper":"https://pith.science/paper/RHVEXAKN","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.07605&json=true","fetch_graph":"https://pith.science/api/pith-number/RHVEXAKNPXGMCBRQSRY7RMLMEW/graph.json","fetch_events":"https://pith.science/api/pith-number/RHVEXAKNPXGMCBRQSRY7RMLMEW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RHVEXAKNPXGMCBRQSRY7RMLMEW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RHVEXAKNPXGMCBRQSRY7RMLMEW/action/storage_attestation","attest_author":"https://pith.science/pith/RHVEXAKNPXGMCBRQSRY7RMLMEW/action/author_attestation","sign_citation":"https://pith.science/pith/RHVEXAKNPXGMCBRQSRY7RMLMEW/action/citation_signature","submit_replication":"https://pith.science/pith/RHVEXAKNPXGMCBRQSRY7RMLMEW/action/replication_record"}},"created_at":"2026-07-05T10:28:08.028557+00:00","updated_at":"2026-07-05T10:28:08.028557+00:00"}