{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:YZAUWJXZVFEKVVG7AQYR4WEMDK","short_pith_number":"pith:YZAUWJXZ","schema_version":"1.0","canonical_sha256":"c6414b26f9a948aad4df04311e588c1a85899aa15d50763e30a6a010f9de0ba8","source":{"kind":"arxiv","id":"2510.04547","version":5},"attestation_state":"computed","paper":{"title":"Activation Quantization of Vision Encoders Needs Prefixing Registers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Jaeho Lee, Jinho Kim, Kyuyeun Kim, SeungHyeon Kim, Taesun Yeom, Wonpyo Park","submitted_at":"2025-10-06T07:27:46Z","abstract_excerpt":"Large pretrained vision encoders are central to multimodal intelligence, powering applications from on-device vision processing to vision-language models. Since these applications often demand real-time processing of massive visual data, reducing the inference cost of vision encoders is critical. Quantization offers a practical path, but it remains challenging even at 8-bit precision due to so-called outliers. In this work, we propose $\\textit{RegCache}$, a training-free algorithm that mitigates outliers in large-scale pretrained vision encoders and serves as a plug-in module that can be appli"},"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":"2510.04547","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-10-06T07:27:46Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"193fd617bf635462d86b82f3ecdefac9fa8ad9e783fc60923ec0b65e8d1e68bf","abstract_canon_sha256":"62912d4809566b4f84a6ea28405d6e8fae5cae3bb9d0726f4cb27bb60b08b933"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-09T01:19:37.892163Z","signature_b64":"UmJ1eEH1EREog9YtblraV6QiuyZcYNkVAk7pqq7KS1aTFl9oRT6VDbZVsBxRTeHv+aj+8eNARZ2/0ZoJk9yNCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c6414b26f9a948aad4df04311e588c1a85899aa15d50763e30a6a010f9de0ba8","last_reissued_at":"2026-07-09T01:19:37.891680Z","signature_status":"signed_v1","first_computed_at":"2026-07-09T01:19:37.891680Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Activation Quantization of Vision Encoders Needs Prefixing Registers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Jaeho Lee, Jinho Kim, Kyuyeun Kim, SeungHyeon Kim, Taesun Yeom, Wonpyo Park","submitted_at":"2025-10-06T07:27:46Z","abstract_excerpt":"Large pretrained vision encoders are central to multimodal intelligence, powering applications from on-device vision processing to vision-language models. Since these applications often demand real-time processing of massive visual data, reducing the inference cost of vision encoders is critical. Quantization offers a practical path, but it remains challenging even at 8-bit precision due to so-called outliers. In this work, we propose $\\textit{RegCache}$, a training-free algorithm that mitigates outliers in large-scale pretrained vision encoders and serves as a plug-in module that can be appli"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.04547","kind":"arxiv","version":5},"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/2510.04547/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":"2510.04547","created_at":"2026-07-09T01:19:37.891736+00:00"},{"alias_kind":"arxiv_version","alias_value":"2510.04547v5","created_at":"2026-07-09T01:19:37.891736+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.04547","created_at":"2026-07-09T01:19:37.891736+00:00"},{"alias_kind":"pith_short_12","alias_value":"YZAUWJXZVFEK","created_at":"2026-07-09T01:19:37.891736+00:00"},{"alias_kind":"pith_short_16","alias_value":"YZAUWJXZVFEKVVG7","created_at":"2026-07-09T01:19:37.891736+00:00"},{"alias_kind":"pith_short_8","alias_value":"YZAUWJXZ","created_at":"2026-07-09T01:19:37.891736+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.20937","citing_title":"Sink-Token-Aware Pruning for Fine-Grained Video Understanding in Efficient Video LLMs","ref_index":21,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YZAUWJXZVFEKVVG7AQYR4WEMDK","json":"https://pith.science/pith/YZAUWJXZVFEKVVG7AQYR4WEMDK.json","graph_json":"https://pith.science/api/pith-number/YZAUWJXZVFEKVVG7AQYR4WEMDK/graph.json","events_json":"https://pith.science/api/pith-number/YZAUWJXZVFEKVVG7AQYR4WEMDK/events.json","paper":"https://pith.science/paper/YZAUWJXZ"},"agent_actions":{"view_html":"https://pith.science/pith/YZAUWJXZVFEKVVG7AQYR4WEMDK","download_json":"https://pith.science/pith/YZAUWJXZVFEKVVG7AQYR4WEMDK.json","view_paper":"https://pith.science/paper/YZAUWJXZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2510.04547&json=true","fetch_graph":"https://pith.science/api/pith-number/YZAUWJXZVFEKVVG7AQYR4WEMDK/graph.json","fetch_events":"https://pith.science/api/pith-number/YZAUWJXZVFEKVVG7AQYR4WEMDK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YZAUWJXZVFEKVVG7AQYR4WEMDK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YZAUWJXZVFEKVVG7AQYR4WEMDK/action/storage_attestation","attest_author":"https://pith.science/pith/YZAUWJXZVFEKVVG7AQYR4WEMDK/action/author_attestation","sign_citation":"https://pith.science/pith/YZAUWJXZVFEKVVG7AQYR4WEMDK/action/citation_signature","submit_replication":"https://pith.science/pith/YZAUWJXZVFEKVVG7AQYR4WEMDK/action/replication_record"}},"created_at":"2026-07-09T01:19:37.891736+00:00","updated_at":"2026-07-09T01:19:37.891736+00:00"}