{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:63QS6NZHBNB2J6ZF5GA4QFYUVE","short_pith_number":"pith:63QS6NZH","canonical_record":{"source":{"id":"2604.16502","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-04-14T14:36:53Z","cross_cats_sorted":[],"title_canon_sha256":"f11f8d27992ce8b86e9a3e4ccc015133a30ff5f475b9e36ec9c7f741a984560f","abstract_canon_sha256":"a8f0eaf27c35ec2b67f410da0c14fece81a52783ea0040d90adc5720abf044a4"},"schema_version":"1.0"},"canonical_sha256":"f6e12f37270b43a4fb25e981c81714a92dff83e212367a7576c332752d6cb2ed","source":{"kind":"arxiv","id":"2604.16502","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.16502","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"arxiv_version","alias_value":"2604.16502v2","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.16502","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_12","alias_value":"63QS6NZHBNB2","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_16","alias_value":"63QS6NZHBNB2J6ZF","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_8","alias_value":"63QS6NZH","created_at":"2026-06-05T01:14:38Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:63QS6NZHBNB2J6ZF5GA4QFYUVE","target":"record","payload":{"canonical_record":{"source":{"id":"2604.16502","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-04-14T14:36:53Z","cross_cats_sorted":[],"title_canon_sha256":"f11f8d27992ce8b86e9a3e4ccc015133a30ff5f475b9e36ec9c7f741a984560f","abstract_canon_sha256":"a8f0eaf27c35ec2b67f410da0c14fece81a52783ea0040d90adc5720abf044a4"},"schema_version":"1.0"},"canonical_sha256":"f6e12f37270b43a4fb25e981c81714a92dff83e212367a7576c332752d6cb2ed","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T01:14:38.873214Z","signature_b64":"KvaY9S8UpYZlbdqAOE2YeShUNZN6yNwlZmBO/dS2YaWVe1oLLGJp7NlrEulpoHZvysaPkJ5WOMLQvWJgNsNRDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f6e12f37270b43a4fb25e981c81714a92dff83e212367a7576c332752d6cb2ed","last_reissued_at":"2026-06-05T01:14:38.872545Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T01:14:38.872545Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.16502","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-05T01:14:38Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7w6349erFjX7ijE12N4LB3QXzDLcXEmFyvAG1835aY0KwicB8H7caq39fwboE6NjLplh+GUU4xvtPmHFrQTbDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-14T18:22:43.952180Z"},"content_sha256":"6f9a90310180f2c3f8cc8244b9f3080d441bedeb2725e91677943781d4dc7e17","schema_version":"1.0","event_id":"sha256:6f9a90310180f2c3f8cc8244b9f3080d441bedeb2725e91677943781d4dc7e17"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:63QS6NZHBNB2J6ZF5GA4QFYUVE","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Topology-Aware Layer Pruning for Large Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Persistent homology on layer point clouds guides pruning to retain critical transitions in vision-language models.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Caiyan Qin, Chaoning Zhang, Jewon Lee, Jiaquan Zhang, Jiarong Mo, Kuien Liu, Pengcheng Zheng, Qigan Sun, Tae-Ho Kim, Tianyu Li, Wang Liu, Yang Yang, Ya Wen","submitted_at":"2026-04-14T14:36:53Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated strong capabilities in natural language understanding and reasoning, while recent extensions that incorporate visual inputs enable them to process multimodal information. Despite these advances, Large Vision-Language Models (LVLMs) incur substantial computational and memory costs, hindering deployment in resource-constrained scenarios. Existing layer pruning methods typically rely on local similarity metrics or static proxy signals, failing to capture the global and dynamic evolution of representations across model depth, which often leads to the "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Extensive experiments on diverse multimodal benchmarks demonstrate that the proposed framework consistently outperforms existing pruning methods across a wide range of sparsity ratios.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That representing layer-wise hidden states as point clouds and quantifying their evolution with simplicial complexes and zigzag persistent homology accurately captures global and dynamic representational transitions better than local similarity metrics or static proxies.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A topology-aware pruning framework models layer representation evolution in LVLMs via simplicial complexes and zigzag persistent homology to enable adaptive removal of layers while outperforming existing methods on multimodal benchmarks.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Persistent homology on layer point clouds guides pruning to retain critical transitions in vision-language models.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"96dde8818129299d44f645034e6161c4932b5647ce0945ba658afd96f768282a"},"source":{"id":"2604.16502","kind":"arxiv","version":2},"verdict":{"id":"f848ebd7-71ea-400d-b727-f960715d85a4","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T16:16:10.385465Z","strongest_claim":"Extensive experiments on diverse multimodal benchmarks demonstrate that the proposed framework consistently outperforms existing pruning methods across a wide range of sparsity ratios.","one_line_summary":"A topology-aware pruning framework models layer representation evolution in LVLMs via simplicial complexes and zigzag persistent homology to enable adaptive removal of layers while outperforming existing methods on multimodal benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That representing layer-wise hidden states as point clouds and quantifying their evolution with simplicial complexes and zigzag persistent homology accurately captures global and dynamic representational transitions better than local similarity metrics or static proxies.","pith_extraction_headline":"Persistent homology on layer point clouds guides pruning to retain critical transitions in vision-language models."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.16502/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":"f848ebd7-71ea-400d-b727-f960715d85a4"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-05T01:14:38Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"m0hESPD1rt6qTZ/6hFZBrynIy1kafpiqLzm/8pRB8FXE6BwVSXo2q6QC4wEzLf1tQZ/QzGimKBIucl5mbRa3AA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-14T18:22:43.952817Z"},"content_sha256":"2a29deed9e0701d17aeb963366e85b707b97ac8131c3a89ec1a8a3e0c0c6497f","schema_version":"1.0","event_id":"sha256:2a29deed9e0701d17aeb963366e85b707b97ac8131c3a89ec1a8a3e0c0c6497f"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/63QS6NZHBNB2J6ZF5GA4QFYUVE/bundle.json","state_url":"https://pith.science/pith/63QS6NZHBNB2J6ZF5GA4QFYUVE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/63QS6NZHBNB2J6ZF5GA4QFYUVE/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":"2026-08-14T18:22:43Z","links":{"resolver":"https://pith.science/pith/63QS6NZHBNB2J6ZF5GA4QFYUVE","bundle":"https://pith.science/pith/63QS6NZHBNB2J6ZF5GA4QFYUVE/bundle.json","state":"https://pith.science/pith/63QS6NZHBNB2J6ZF5GA4QFYUVE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/63QS6NZHBNB2J6ZF5GA4QFYUVE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:63QS6NZHBNB2J6ZF5GA4QFYUVE","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":"a8f0eaf27c35ec2b67f410da0c14fece81a52783ea0040d90adc5720abf044a4","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-04-14T14:36:53Z","title_canon_sha256":"f11f8d27992ce8b86e9a3e4ccc015133a30ff5f475b9e36ec9c7f741a984560f"},"schema_version":"1.0","source":{"id":"2604.16502","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.16502","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"arxiv_version","alias_value":"2604.16502v2","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.16502","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_12","alias_value":"63QS6NZHBNB2","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_16","alias_value":"63QS6NZHBNB2J6ZF","created_at":"2026-06-05T01:14:38Z"},{"alias_kind":"pith_short_8","alias_value":"63QS6NZH","created_at":"2026-06-05T01:14:38Z"}],"graph_snapshots":[{"event_id":"sha256:2a29deed9e0701d17aeb963366e85b707b97ac8131c3a89ec1a8a3e0c0c6497f","target":"graph","created_at":"2026-06-05T01:14:38Z","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":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Extensive experiments on diverse multimodal benchmarks demonstrate that the proposed framework consistently outperforms existing pruning methods across a wide range of sparsity ratios."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That representing layer-wise hidden states as point clouds and quantifying their evolution with simplicial complexes and zigzag persistent homology accurately captures global and dynamic representational transitions better than local similarity metrics or static proxies."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"A topology-aware pruning framework models layer representation evolution in LVLMs via simplicial complexes and zigzag persistent homology to enable adaptive removal of layers while outperforming existing methods on multimodal benchmarks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Persistent homology on layer point clouds guides pruning to retain critical transitions in vision-language models."}],"snapshot_sha256":"96dde8818129299d44f645034e6161c4932b5647ce0945ba658afd96f768282a"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2604.16502/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) have demonstrated strong capabilities in natural language understanding and reasoning, while recent extensions that incorporate visual inputs enable them to process multimodal information. Despite these advances, Large Vision-Language Models (LVLMs) incur substantial computational and memory costs, hindering deployment in resource-constrained scenarios. Existing layer pruning methods typically rely on local similarity metrics or static proxy signals, failing to capture the global and dynamic evolution of representations across model depth, which often leads to the ","authors_text":"Caiyan Qin, Chaoning Zhang, Jewon Lee, Jiaquan Zhang, Jiarong Mo, Kuien Liu, Pengcheng Zheng, Qigan Sun, Tae-Ho Kim, Tianyu Li, Wang Liu, Yang Yang, Ya Wen","cross_cats":[],"headline":"Persistent homology on layer point clouds guides pruning to retain critical transitions in vision-language models.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-04-14T14:36:53Z","title":"Topology-Aware Layer Pruning for Large Vision-Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.16502","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-10T16:16:10.385465Z","id":"f848ebd7-71ea-400d-b727-f960715d85a4","model_set":{"reader":"grok-4.3"},"one_line_summary":"A topology-aware pruning framework models layer representation evolution in LVLMs via simplicial complexes and zigzag persistent homology to enable adaptive removal of layers while outperforming existing methods on multimodal benchmarks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Persistent homology on layer point clouds guides pruning to retain critical transitions in vision-language models.","strongest_claim":"Extensive experiments on diverse multimodal benchmarks demonstrate that the proposed framework consistently outperforms existing pruning methods across a wide range of sparsity ratios.","weakest_assumption":"That representing layer-wise hidden states as point clouds and quantifying their evolution with simplicial complexes and zigzag persistent homology accurately captures global and dynamic representational transitions better than local similarity metrics or static proxies."}},"verdict_id":"f848ebd7-71ea-400d-b727-f960715d85a4"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:6f9a90310180f2c3f8cc8244b9f3080d441bedeb2725e91677943781d4dc7e17","target":"record","created_at":"2026-06-05T01:14:38Z","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":"a8f0eaf27c35ec2b67f410da0c14fece81a52783ea0040d90adc5720abf044a4","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-04-14T14:36:53Z","title_canon_sha256":"f11f8d27992ce8b86e9a3e4ccc015133a30ff5f475b9e36ec9c7f741a984560f"},"schema_version":"1.0","source":{"id":"2604.16502","kind":"arxiv","version":2}},"canonical_sha256":"f6e12f37270b43a4fb25e981c81714a92dff83e212367a7576c332752d6cb2ed","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f6e12f37270b43a4fb25e981c81714a92dff83e212367a7576c332752d6cb2ed","first_computed_at":"2026-06-05T01:14:38.872545Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-05T01:14:38.872545Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KvaY9S8UpYZlbdqAOE2YeShUNZN6yNwlZmBO/dS2YaWVe1oLLGJp7NlrEulpoHZvysaPkJ5WOMLQvWJgNsNRDQ==","signature_status":"signed_v1","signed_at":"2026-06-05T01:14:38.873214Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.16502","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6f9a90310180f2c3f8cc8244b9f3080d441bedeb2725e91677943781d4dc7e17","sha256:2a29deed9e0701d17aeb963366e85b707b97ac8131c3a89ec1a8a3e0c0c6497f"],"state_sha256":"a2f1186ff7129dcd6bc3cee8f6f9b5ab6ff869e16c8ac986fe2b4c4a056f8869"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"JAuwnh4hnx4xivCM8ETjKvPWMjoo2s9NOpwRYfeejtnM31URZiRzasourtg1Q9/lMfM3iVuCwASr19MjHd8cBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-14T18:22:43.957396Z","bundle_sha256":"ffc94481e6cfc16ffa32d84b03e97c23f271df569f0c1b901531a36cb53cc2d4"}}