{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:34EWLCNOG6XGDQCH44KUM2E7YA","short_pith_number":"pith:34EWLCNO","schema_version":"1.0","canonical_sha256":"df096589ae37ae61c047e71546689fc036e74e941d7af41e94f3408af1d1c0c4","source":{"kind":"arxiv","id":"2605.18795","version":1},"attestation_state":"computed","paper":{"title":"HELLoRA: Hot Experts Layer-Level Low-Rank Adaptation for Mixture-of-Experts Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jia Wei, Longxiang Wang, Ping Chen, Qianyang li, Shaoxun Wang, Yancheng Pan, Zhonghao Zhang, Ziyi Qiu","submitted_at":"2026-05-11T06:43:14Z","abstract_excerpt":"Low-Rank Adaptation (LoRA) dominates parameter-efficient fine-tuning of large language models, yet most variants target dense architectures. Mixture-of-Experts (MoE) models scale parameters at near-constant per-token compute, and their sparse activation patterns create untapped opportunities for more efficient adaptation. We propose Hot-Experts Layer-level Low-Rank Adaptation (HELLoRA), which attaches LoRA modules only to the most frequently activated experts at each layer. This simple mechanism reduces trainable parameters and adapter-induced FLOPs while improving downstream performance, an e"},"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":"2605.18795","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-11T06:43:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"911aeb2b49ba4b07e61a4684d843ec95809a106ebe8f49cf010a5e051331f386","abstract_canon_sha256":"46106e97312f102580af34a0d2a3eee1aad85ffff811775eebe19f7057db4362"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:06:22.816747Z","signature_b64":"gxCcKnGydZyE6qXgd5W6fbFjw5gxos/V5LurC4df7sRNvK+uiyNklBF+65rie2ppoQJ3biRPrxkm5HW4w5ODCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"df096589ae37ae61c047e71546689fc036e74e941d7af41e94f3408af1d1c0c4","last_reissued_at":"2026-05-20T00:06:22.815746Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:06:22.815746Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HELLoRA: Hot Experts Layer-Level Low-Rank Adaptation for Mixture-of-Experts Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Jia Wei, Longxiang Wang, Ping Chen, Qianyang li, Shaoxun Wang, Yancheng Pan, Zhonghao Zhang, Ziyi Qiu","submitted_at":"2026-05-11T06:43:14Z","abstract_excerpt":"Low-Rank Adaptation (LoRA) dominates parameter-efficient fine-tuning of large language models, yet most variants target dense architectures. Mixture-of-Experts (MoE) models scale parameters at near-constant per-token compute, and their sparse activation patterns create untapped opportunities for more efficient adaptation. We propose Hot-Experts Layer-level Low-Rank Adaptation (HELLoRA), which attaches LoRA modules only to the most frequently activated experts at each layer. This simple mechanism reduces trainable parameters and adapter-induced FLOPs while improving downstream performance, an e"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.18795","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/2605.18795/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":"2605.18795","created_at":"2026-05-20T00:06:22.815899+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.18795v1","created_at":"2026-05-20T00:06:22.815899+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.18795","created_at":"2026-05-20T00:06:22.815899+00:00"},{"alias_kind":"pith_short_12","alias_value":"34EWLCNOG6XG","created_at":"2026-05-20T00:06:22.815899+00:00"},{"alias_kind":"pith_short_16","alias_value":"34EWLCNOG6XGDQCH","created_at":"2026-05-20T00:06:22.815899+00:00"},{"alias_kind":"pith_short_8","alias_value":"34EWLCNO","created_at":"2026-05-20T00:06:22.815899+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/34EWLCNOG6XGDQCH44KUM2E7YA","json":"https://pith.science/pith/34EWLCNOG6XGDQCH44KUM2E7YA.json","graph_json":"https://pith.science/api/pith-number/34EWLCNOG6XGDQCH44KUM2E7YA/graph.json","events_json":"https://pith.science/api/pith-number/34EWLCNOG6XGDQCH44KUM2E7YA/events.json","paper":"https://pith.science/paper/34EWLCNO"},"agent_actions":{"view_html":"https://pith.science/pith/34EWLCNOG6XGDQCH44KUM2E7YA","download_json":"https://pith.science/pith/34EWLCNOG6XGDQCH44KUM2E7YA.json","view_paper":"https://pith.science/paper/34EWLCNO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.18795&json=true","fetch_graph":"https://pith.science/api/pith-number/34EWLCNOG6XGDQCH44KUM2E7YA/graph.json","fetch_events":"https://pith.science/api/pith-number/34EWLCNOG6XGDQCH44KUM2E7YA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/34EWLCNOG6XGDQCH44KUM2E7YA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/34EWLCNOG6XGDQCH44KUM2E7YA/action/storage_attestation","attest_author":"https://pith.science/pith/34EWLCNOG6XGDQCH44KUM2E7YA/action/author_attestation","sign_citation":"https://pith.science/pith/34EWLCNOG6XGDQCH44KUM2E7YA/action/citation_signature","submit_replication":"https://pith.science/pith/34EWLCNOG6XGDQCH44KUM2E7YA/action/replication_record"}},"created_at":"2026-05-20T00:06:22.815899+00:00","updated_at":"2026-05-20T00:06:22.815899+00:00"}