{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TXM5QQK5GEIOQ3G5M4XR662DK6","short_pith_number":"pith:TXM5QQK5","schema_version":"1.0","canonical_sha256":"9dd9d8415d3110e86cdd672f1f7b4357baeb34527cd9b01e972e31e7139fe98f","source":{"kind":"arxiv","id":"2606.09886","version":1},"attestation_state":"computed","paper":{"title":"SHAPE: Coalition-Aware Expert Pruning for Sparse Mixture-of-Experts LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Yuhao Zhang","submitted_at":"2026-06-03T08:41:48Z","abstract_excerpt":"Sparse Mixture-of-Experts (MoE) large language models achieve strong quality with low per-token compute, yet their deployment is often limited by the memory wall: the full expert pool must remain resident to support token-dependent routing. Expert pruning is a direct remedy, but prior criteria typically score experts independently and overlook that MoE inference is inherently \\emph{coalitional}, where outputs arise from routed top-$k$ expert combinations. We propose \\textbf{SHAPE}, a task-driven pruning framework that explicitly models \\emph{intra-layer} expert cooperation. SHAPE formulates ro"},"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":"2606.09886","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T08:41:48Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"68e06b75b6fe967738e71511c3ee568f7e0c67c6076a50790d7e22febef45c7d","abstract_canon_sha256":"b7070185db30b180a34443d2a1c29ca1c9ffd9df90d24a1ef87d814c1d6f4316"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-10T00:08:31.283805Z","signature_b64":"aU71dm0HZswu+0rovFCTYuqD4a9RlBs/pp9L/lB2ELKfOuMOIY6oag9FaHnFqe0veWuyBJ/Sqn7ZQy9QQNu6AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9dd9d8415d3110e86cdd672f1f7b4357baeb34527cd9b01e972e31e7139fe98f","last_reissued_at":"2026-06-10T00:08:31.282940Z","signature_status":"signed_v1","first_computed_at":"2026-06-10T00:08:31.282940Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SHAPE: Coalition-Aware Expert Pruning for Sparse Mixture-of-Experts LLMs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Yuhao Zhang","submitted_at":"2026-06-03T08:41:48Z","abstract_excerpt":"Sparse Mixture-of-Experts (MoE) large language models achieve strong quality with low per-token compute, yet their deployment is often limited by the memory wall: the full expert pool must remain resident to support token-dependent routing. Expert pruning is a direct remedy, but prior criteria typically score experts independently and overlook that MoE inference is inherently \\emph{coalitional}, where outputs arise from routed top-$k$ expert combinations. We propose \\textbf{SHAPE}, a task-driven pruning framework that explicitly models \\emph{intra-layer} expert cooperation. SHAPE formulates ro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.09886","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/2606.09886/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":"2606.09886","created_at":"2026-06-10T00:08:31.283080+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.09886v1","created_at":"2026-06-10T00:08:31.283080+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.09886","created_at":"2026-06-10T00:08:31.283080+00:00"},{"alias_kind":"pith_short_12","alias_value":"TXM5QQK5GEIO","created_at":"2026-06-10T00:08:31.283080+00:00"},{"alias_kind":"pith_short_16","alias_value":"TXM5QQK5GEIOQ3G5","created_at":"2026-06-10T00:08:31.283080+00:00"},{"alias_kind":"pith_short_8","alias_value":"TXM5QQK5","created_at":"2026-06-10T00:08:31.283080+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/TXM5QQK5GEIOQ3G5M4XR662DK6","json":"https://pith.science/pith/TXM5QQK5GEIOQ3G5M4XR662DK6.json","graph_json":"https://pith.science/api/pith-number/TXM5QQK5GEIOQ3G5M4XR662DK6/graph.json","events_json":"https://pith.science/api/pith-number/TXM5QQK5GEIOQ3G5M4XR662DK6/events.json","paper":"https://pith.science/paper/TXM5QQK5"},"agent_actions":{"view_html":"https://pith.science/pith/TXM5QQK5GEIOQ3G5M4XR662DK6","download_json":"https://pith.science/pith/TXM5QQK5GEIOQ3G5M4XR662DK6.json","view_paper":"https://pith.science/paper/TXM5QQK5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.09886&json=true","fetch_graph":"https://pith.science/api/pith-number/TXM5QQK5GEIOQ3G5M4XR662DK6/graph.json","fetch_events":"https://pith.science/api/pith-number/TXM5QQK5GEIOQ3G5M4XR662DK6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TXM5QQK5GEIOQ3G5M4XR662DK6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TXM5QQK5GEIOQ3G5M4XR662DK6/action/storage_attestation","attest_author":"https://pith.science/pith/TXM5QQK5GEIOQ3G5M4XR662DK6/action/author_attestation","sign_citation":"https://pith.science/pith/TXM5QQK5GEIOQ3G5M4XR662DK6/action/citation_signature","submit_replication":"https://pith.science/pith/TXM5QQK5GEIOQ3G5M4XR662DK6/action/replication_record"}},"created_at":"2026-06-10T00:08:31.283080+00:00","updated_at":"2026-06-10T00:08:31.283080+00:00"}