{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:LV63QY2TKBSOIDCVTAQSFH5YHE","short_pith_number":"pith:LV63QY2T","canonical_record":{"source":{"id":"2501.15021","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-25T02:01:56Z","cross_cats_sorted":[],"title_canon_sha256":"d570ee20808418c1fb1d8e999b59fc63ed1d719cabad326e3b14575d1dbd2af5","abstract_canon_sha256":"8e3f14384e93ddc581b81c45aaa52b671791974f776e06e77aa3af85f9117162"},"schema_version":"1.0"},"canonical_sha256":"5d7db863535064e40c559821229fb8392ba401aa7e3e879149b776fa924f7a82","source":{"kind":"arxiv","id":"2501.15021","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.15021","created_at":"2026-07-05T10:05:22Z"},{"alias_kind":"arxiv_version","alias_value":"2501.15021v1","created_at":"2026-07-05T10:05:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.15021","created_at":"2026-07-05T10:05:22Z"},{"alias_kind":"pith_short_12","alias_value":"LV63QY2TKBSO","created_at":"2026-07-05T10:05:22Z"},{"alias_kind":"pith_short_16","alias_value":"LV63QY2TKBSOIDCV","created_at":"2026-07-05T10:05:22Z"},{"alias_kind":"pith_short_8","alias_value":"LV63QY2T","created_at":"2026-07-05T10:05:22Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:LV63QY2TKBSOIDCVTAQSFH5YHE","target":"record","payload":{"canonical_record":{"source":{"id":"2501.15021","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-25T02:01:56Z","cross_cats_sorted":[],"title_canon_sha256":"d570ee20808418c1fb1d8e999b59fc63ed1d719cabad326e3b14575d1dbd2af5","abstract_canon_sha256":"8e3f14384e93ddc581b81c45aaa52b671791974f776e06e77aa3af85f9117162"},"schema_version":"1.0"},"canonical_sha256":"5d7db863535064e40c559821229fb8392ba401aa7e3e879149b776fa924f7a82","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:05:22.330911Z","signature_b64":"NM9KMufrEeL216p1M1+Z3ulDqJFokit/nfCVGlapzGyGSz5e+kNcKV//slO2xZA9AYIma/WnI9Te8nobr/KgDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5d7db863535064e40c559821229fb8392ba401aa7e3e879149b776fa924f7a82","last_reissued_at":"2026-07-05T10:05:22.330502Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:05:22.330502Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2501.15021","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:05:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nqfJrgiPjMHEsUSrFTJwROwtDf+HBXbqEelfvsQjrQt2vEinGa9YPLZZqle0FUgDP2V4k/+dysg+o0bxkd7OCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T09:23:27.761589Z"},"content_sha256":"068330e3a362e8d6b466dba841c053ab82c6467c6f663baee493f8dc20c55d13","schema_version":"1.0","event_id":"sha256:068330e3a362e8d6b466dba841c053ab82c6467c6f663baee493f8dc20c55d13"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:LV63QY2TKBSOIDCVTAQSFH5YHE","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AKVQ-VL: Attention-Aware KV Cache Adaptive 2-Bit Quantization for Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hanyu Wei, Huangqi Yu, Kehong Yuan, Linge Li, Wang Shen, Zhe Chen, Zunhai Su","submitted_at":"2025-01-25T02:01:56Z","abstract_excerpt":"Vision-language models (VLMs) show remarkable performance in multimodal tasks. However, excessively long multimodal inputs lead to oversized Key-Value (KV) caches, resulting in significant memory consumption and I/O bottlenecks. Previous KV quantization methods for Large Language Models (LLMs) may alleviate these issues but overlook the attention saliency differences of multimodal tokens, resulting in suboptimal performance. In this paper, we investigate the attention-aware token saliency patterns in VLM and propose AKVQ-VL. AKVQ-VL leverages the proposed Text-Salient Attention (TSA) and Pivot"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.15021","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/2501.15021/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:05:22Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZLkV2oEc8H8at0BeXx2SvCl78qd9pSc/LP09icG8bFvCiHKbS1iNkgxaPov765F7V8WdjSJVsrUG4JTyRIx/Dw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T09:23:27.762094Z"},"content_sha256":"69dc487789c8767a7d88199522c80b9c6f30e26f54995e5e7ab448e99688802a","schema_version":"1.0","event_id":"sha256:69dc487789c8767a7d88199522c80b9c6f30e26f54995e5e7ab448e99688802a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LV63QY2TKBSOIDCVTAQSFH5YHE/bundle.json","state_url":"https://pith.science/pith/LV63QY2TKBSOIDCVTAQSFH5YHE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LV63QY2TKBSOIDCVTAQSFH5YHE/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-06T09:23:27Z","links":{"resolver":"https://pith.science/pith/LV63QY2TKBSOIDCVTAQSFH5YHE","bundle":"https://pith.science/pith/LV63QY2TKBSOIDCVTAQSFH5YHE/bundle.json","state":"https://pith.science/pith/LV63QY2TKBSOIDCVTAQSFH5YHE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LV63QY2TKBSOIDCVTAQSFH5YHE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:LV63QY2TKBSOIDCVTAQSFH5YHE","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"8e3f14384e93ddc581b81c45aaa52b671791974f776e06e77aa3af85f9117162","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-25T02:01:56Z","title_canon_sha256":"d570ee20808418c1fb1d8e999b59fc63ed1d719cabad326e3b14575d1dbd2af5"},"schema_version":"1.0","source":{"id":"2501.15021","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.15021","created_at":"2026-07-05T10:05:22Z"},{"alias_kind":"arxiv_version","alias_value":"2501.15021v1","created_at":"2026-07-05T10:05:22Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.15021","created_at":"2026-07-05T10:05:22Z"},{"alias_kind":"pith_short_12","alias_value":"LV63QY2TKBSO","created_at":"2026-07-05T10:05:22Z"},{"alias_kind":"pith_short_16","alias_value":"LV63QY2TKBSOIDCV","created_at":"2026-07-05T10:05:22Z"},{"alias_kind":"pith_short_8","alias_value":"LV63QY2T","created_at":"2026-07-05T10:05:22Z"}],"graph_snapshots":[{"event_id":"sha256:69dc487789c8767a7d88199522c80b9c6f30e26f54995e5e7ab448e99688802a","target":"graph","created_at":"2026-07-05T10:05:22Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2501.15021/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Vision-language models (VLMs) show remarkable performance in multimodal tasks. However, excessively long multimodal inputs lead to oversized Key-Value (KV) caches, resulting in significant memory consumption and I/O bottlenecks. Previous KV quantization methods for Large Language Models (LLMs) may alleviate these issues but overlook the attention saliency differences of multimodal tokens, resulting in suboptimal performance. In this paper, we investigate the attention-aware token saliency patterns in VLM and propose AKVQ-VL. AKVQ-VL leverages the proposed Text-Salient Attention (TSA) and Pivot","authors_text":"Hanyu Wei, Huangqi Yu, Kehong Yuan, Linge Li, Wang Shen, Zhe Chen, Zunhai Su","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-25T02:01:56Z","title":"AKVQ-VL: Attention-Aware KV Cache Adaptive 2-Bit Quantization for Vision-Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.15021","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:068330e3a362e8d6b466dba841c053ab82c6467c6f663baee493f8dc20c55d13","target":"record","created_at":"2026-07-05T10:05:22Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"8e3f14384e93ddc581b81c45aaa52b671791974f776e06e77aa3af85f9117162","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-25T02:01:56Z","title_canon_sha256":"d570ee20808418c1fb1d8e999b59fc63ed1d719cabad326e3b14575d1dbd2af5"},"schema_version":"1.0","source":{"id":"2501.15021","kind":"arxiv","version":1}},"canonical_sha256":"5d7db863535064e40c559821229fb8392ba401aa7e3e879149b776fa924f7a82","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5d7db863535064e40c559821229fb8392ba401aa7e3e879149b776fa924f7a82","first_computed_at":"2026-07-05T10:05:22.330502Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:05:22.330502Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"NM9KMufrEeL216p1M1+Z3ulDqJFokit/nfCVGlapzGyGSz5e+kNcKV//slO2xZA9AYIma/WnI9Te8nobr/KgDw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:05:22.330911Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.15021","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:068330e3a362e8d6b466dba841c053ab82c6467c6f663baee493f8dc20c55d13","sha256:69dc487789c8767a7d88199522c80b9c6f30e26f54995e5e7ab448e99688802a"],"state_sha256":"505947100149c0876298ff8a3088149ba299d5840f904c1354af746b9af691fe"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"M58tx9w0PVSJdsMzs2Zv7aZRw3rT9dbrKiafuF1VKmljBBMXat6prJVuDbRvanVMgNupResUYCRiYViVPyqaAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T09:23:27.765893Z","bundle_sha256":"5ea92df064f5b285eae07dbea77c19c65bdb012bd701adf5344b307e9cf5a8d0"}}