{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:3HTPYKFUFRX7V6YIKACQGUIUJM","short_pith_number":"pith:3HTPYKFU","schema_version":"1.0","canonical_sha256":"d9e6fc28b42c6ffafb0850050351144b1739a274b7d89c829aef7312f5fe613f","source":{"kind":"arxiv","id":"2606.17118","version":1},"attestation_state":"computed","paper":{"title":"MODE: Modality-Decomposed Expert-Level Mixed-Precision Quantization for MoE Multimodal LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chuangyi Li, Gang Li, Jian Cheng, Jing Liu, Nanxin Zeng, Peisong Wang, Qinghao Hu, Shiqiang Lang, Tao Liu, Yuanteng Chen, Yuantian Shao, Zhilei Liu","submitted_at":"2026-06-15T10:59:11Z","abstract_excerpt":"Mixture-of-Experts Multimodal Large Language Models (MoE-MLLMs) offer remarkable performance but incur prohibitive GPU memory costs, making compression essential. Among PTQ methods, expert-level mixed-precision quantization has proven effective for MoE-LLMs, yet suffers notable degradation on MoE-MLLMs due to two overlooked biases in expert importance estimation. (1) At the cross-modal level, the numerical dominance of vision tokens causes expert selection frequency to be dominated by vision tokens, masking experts that are critical to the text modality; (2) at the intra-vision level, the larg"},"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":"2606.17118","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-06-15T10:59:11Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b8286b3c112a413bcd852056997ec349cf566d124b20ef246a21966b818d1376","abstract_canon_sha256":"2f84fc3ee96921bd215961de5f35cc4836c17f1e6e79492636b900fe8181eef6"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:10:03.599582Z","signature_b64":"3vue7vt/qDb35zKwfjgn7YJcLTP8CteHwcecM/ARwjlNL84VpVltPcogVITtoMjWJu3yw4ISZEeBNgkHwW9HAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d9e6fc28b42c6ffafb0850050351144b1739a274b7d89c829aef7312f5fe613f","last_reissued_at":"2026-06-19T16:10:03.599176Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:10:03.599176Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MODE: Modality-Decomposed Expert-Level Mixed-Precision Quantization for MoE Multimodal LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Chuangyi Li, Gang Li, Jian Cheng, Jing Liu, Nanxin Zeng, Peisong Wang, Qinghao Hu, Shiqiang Lang, Tao Liu, Yuanteng Chen, Yuantian Shao, Zhilei Liu","submitted_at":"2026-06-15T10:59:11Z","abstract_excerpt":"Mixture-of-Experts Multimodal Large Language Models (MoE-MLLMs) offer remarkable performance but incur prohibitive GPU memory costs, making compression essential. Among PTQ methods, expert-level mixed-precision quantization has proven effective for MoE-LLMs, yet suffers notable degradation on MoE-MLLMs due to two overlooked biases in expert importance estimation. (1) At the cross-modal level, the numerical dominance of vision tokens causes expert selection frequency to be dominated by vision tokens, masking experts that are critical to the text modality; (2) at the intra-vision level, the larg"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.17118","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/2606.17118/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":"2606.17118","created_at":"2026-06-19T16:10:03.599235+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.17118v1","created_at":"2026-06-19T16:10:03.599235+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.17118","created_at":"2026-06-19T16:10:03.599235+00:00"},{"alias_kind":"pith_short_12","alias_value":"3HTPYKFUFRX7","created_at":"2026-06-19T16:10:03.599235+00:00"},{"alias_kind":"pith_short_16","alias_value":"3HTPYKFUFRX7V6YI","created_at":"2026-06-19T16:10:03.599235+00:00"},{"alias_kind":"pith_short_8","alias_value":"3HTPYKFU","created_at":"2026-06-19T16:10:03.599235+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/3HTPYKFUFRX7V6YIKACQGUIUJM","json":"https://pith.science/pith/3HTPYKFUFRX7V6YIKACQGUIUJM.json","graph_json":"https://pith.science/api/pith-number/3HTPYKFUFRX7V6YIKACQGUIUJM/graph.json","events_json":"https://pith.science/api/pith-number/3HTPYKFUFRX7V6YIKACQGUIUJM/events.json","paper":"https://pith.science/paper/3HTPYKFU"},"agent_actions":{"view_html":"https://pith.science/pith/3HTPYKFUFRX7V6YIKACQGUIUJM","download_json":"https://pith.science/pith/3HTPYKFUFRX7V6YIKACQGUIUJM.json","view_paper":"https://pith.science/paper/3HTPYKFU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.17118&json=true","fetch_graph":"https://pith.science/api/pith-number/3HTPYKFUFRX7V6YIKACQGUIUJM/graph.json","fetch_events":"https://pith.science/api/pith-number/3HTPYKFUFRX7V6YIKACQGUIUJM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3HTPYKFUFRX7V6YIKACQGUIUJM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3HTPYKFUFRX7V6YIKACQGUIUJM/action/storage_attestation","attest_author":"https://pith.science/pith/3HTPYKFUFRX7V6YIKACQGUIUJM/action/author_attestation","sign_citation":"https://pith.science/pith/3HTPYKFUFRX7V6YIKACQGUIUJM/action/citation_signature","submit_replication":"https://pith.science/pith/3HTPYKFUFRX7V6YIKACQGUIUJM/action/replication_record"}},"created_at":"2026-06-19T16:10:03.599235+00:00","updated_at":"2026-06-19T16:10:03.599235+00:00"}