{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:LWZ6PLF3WNKIGOBDQ4BCWAAS5N","short_pith_number":"pith:LWZ6PLF3","canonical_record":{"source":{"id":"2608.04496","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-08-05T06:32:25Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"abb1979733dbe48e6c9322cc58ac4bc44c2c31973f6f553ede9613e4635e6803","abstract_canon_sha256":"88ed1e84f3bcb58687cf04f4079fe1c179668539b4cb1409199970239e037529"},"schema_version":"1.0"},"canonical_sha256":"5db3e7acbbb35483382387022b0012eb5be66807d8a47934c2641c17e0b7b0de","source":{"kind":"arxiv","id":"2608.04496","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.04496","created_at":"2026-08-06T01:35:42Z"},{"alias_kind":"arxiv_version","alias_value":"2608.04496v1","created_at":"2026-08-06T01:35:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.04496","created_at":"2026-08-06T01:35:42Z"},{"alias_kind":"pith_short_12","alias_value":"LWZ6PLF3WNKI","created_at":"2026-08-06T01:35:42Z"},{"alias_kind":"pith_short_16","alias_value":"LWZ6PLF3WNKIGOBD","created_at":"2026-08-06T01:35:42Z"},{"alias_kind":"pith_short_8","alias_value":"LWZ6PLF3","created_at":"2026-08-06T01:35:42Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:LWZ6PLF3WNKIGOBDQ4BCWAAS5N","target":"record","payload":{"canonical_record":{"source":{"id":"2608.04496","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-08-05T06:32:25Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"abb1979733dbe48e6c9322cc58ac4bc44c2c31973f6f553ede9613e4635e6803","abstract_canon_sha256":"88ed1e84f3bcb58687cf04f4079fe1c179668539b4cb1409199970239e037529"},"schema_version":"1.0"},"canonical_sha256":"5db3e7acbbb35483382387022b0012eb5be66807d8a47934c2641c17e0b7b0de","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-06T01:35:42.858835Z","signature_b64":"uEIHHnUUKnA/ANj4RsXNFxWXPywLED1soBoo5pq8IlHFFn3awBblbRdcBMsVhYEdGK33yEJr96qcXhhhfvhXDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5db3e7acbbb35483382387022b0012eb5be66807d8a47934c2641c17e0b7b0de","last_reissued_at":"2026-08-06T01:35:42.857456Z","signature_status":"signed_v1","first_computed_at":"2026-08-06T01:35:42.857456Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2608.04496","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":"2026-08-06T01:35:42Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"pdHaFrefIgLsklv7cOf9mow0prWVpQnQe2oFT5IzC/w+YZfJVC/EhUJVKTXqRbEWpxrWRZMBKqL0wntSx7GEDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-24T02:09:05.878762Z"},"content_sha256":"0b93fd1bb987ba2cb89c9d6867d680807585e41e343ebdfc8d42597b34becd25","schema_version":"1.0","event_id":"sha256:0b93fd1bb987ba2cb89c9d6867d680807585e41e343ebdfc8d42597b34becd25"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:LWZ6PLF3WNKIGOBDQ4BCWAAS5N","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"DIVE: Dynamic Iterative Visual Evidence Construction for Efficient Vision-Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Chen Zhong, Guangyi Yang, Jiepan Li, Wei He, Xiao An, Zijie Wang","submitted_at":"2026-08-05T06:32:25Z","abstract_excerpt":"Visual inputs in vision-language models (VLMs) are often encoded into substantially longer token sequences than text, making visual tokens a major bottleneck for efficient inference. Abundant recent methods address this bottleneck by scoring token importance and pruning low-scoring tokens in a single pass. However, one-shot scoring is insufficient because a token's prompt-relevant usefulness depends on the evidence already retained. Motivated by this insight, we introduce DIVE (Dynamic Iterative Visual Evidence Construction), a training-free framework that recasts visual-token pruning as dynam"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.04496","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/2608.04496/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":"2026-08-06T01:35:42Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8TDxgkAjCaTtky/H031R7ovlLtyTW0y0cN1KaVEpmBIjWWJZqJ1HaMSJPCPCPsZbgmMd112P38FI4LGCgVtSBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-24T02:09:05.879277Z"},"content_sha256":"bc5f5e9f57980c1db4a0c055fda157e47d451ef5bd72af59e2d8992a505c1ab2","schema_version":"1.0","event_id":"sha256:bc5f5e9f57980c1db4a0c055fda157e47d451ef5bd72af59e2d8992a505c1ab2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LWZ6PLF3WNKIGOBDQ4BCWAAS5N/bundle.json","state_url":"https://pith.science/pith/LWZ6PLF3WNKIGOBDQ4BCWAAS5N/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LWZ6PLF3WNKIGOBDQ4BCWAAS5N/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":"2026-08-24T02:09:05Z","links":{"resolver":"https://pith.science/pith/LWZ6PLF3WNKIGOBDQ4BCWAAS5N","bundle":"https://pith.science/pith/LWZ6PLF3WNKIGOBDQ4BCWAAS5N/bundle.json","state":"https://pith.science/pith/LWZ6PLF3WNKIGOBDQ4BCWAAS5N/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LWZ6PLF3WNKIGOBDQ4BCWAAS5N/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:LWZ6PLF3WNKIGOBDQ4BCWAAS5N","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":"88ed1e84f3bcb58687cf04f4079fe1c179668539b4cb1409199970239e037529","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-08-05T06:32:25Z","title_canon_sha256":"abb1979733dbe48e6c9322cc58ac4bc44c2c31973f6f553ede9613e4635e6803"},"schema_version":"1.0","source":{"id":"2608.04496","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.04496","created_at":"2026-08-06T01:35:42Z"},{"alias_kind":"arxiv_version","alias_value":"2608.04496v1","created_at":"2026-08-06T01:35:42Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.04496","created_at":"2026-08-06T01:35:42Z"},{"alias_kind":"pith_short_12","alias_value":"LWZ6PLF3WNKI","created_at":"2026-08-06T01:35:42Z"},{"alias_kind":"pith_short_16","alias_value":"LWZ6PLF3WNKIGOBD","created_at":"2026-08-06T01:35:42Z"},{"alias_kind":"pith_short_8","alias_value":"LWZ6PLF3","created_at":"2026-08-06T01:35:42Z"}],"graph_snapshots":[{"event_id":"sha256:bc5f5e9f57980c1db4a0c055fda157e47d451ef5bd72af59e2d8992a505c1ab2","target":"graph","created_at":"2026-08-06T01:35:42Z","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/2608.04496/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Visual inputs in vision-language models (VLMs) are often encoded into substantially longer token sequences than text, making visual tokens a major bottleneck for efficient inference. Abundant recent methods address this bottleneck by scoring token importance and pruning low-scoring tokens in a single pass. However, one-shot scoring is insufficient because a token's prompt-relevant usefulness depends on the evidence already retained. Motivated by this insight, we introduce DIVE (Dynamic Iterative Visual Evidence Construction), a training-free framework that recasts visual-token pruning as dynam","authors_text":"Chen Zhong, Guangyi Yang, Jiepan Li, Wei He, Xiao An, Zijie Wang","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-08-05T06:32:25Z","title":"DIVE: Dynamic Iterative Visual Evidence Construction for Efficient Vision-Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.04496","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":"sha256:0b93fd1bb987ba2cb89c9d6867d680807585e41e343ebdfc8d42597b34becd25","target":"record","created_at":"2026-08-06T01:35:42Z","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":"88ed1e84f3bcb58687cf04f4079fe1c179668539b4cb1409199970239e037529","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-08-05T06:32:25Z","title_canon_sha256":"abb1979733dbe48e6c9322cc58ac4bc44c2c31973f6f553ede9613e4635e6803"},"schema_version":"1.0","source":{"id":"2608.04496","kind":"arxiv","version":1}},"canonical_sha256":"5db3e7acbbb35483382387022b0012eb5be66807d8a47934c2641c17e0b7b0de","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5db3e7acbbb35483382387022b0012eb5be66807d8a47934c2641c17e0b7b0de","first_computed_at":"2026-08-06T01:35:42.857456Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-08-06T01:35:42.857456Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"uEIHHnUUKnA/ANj4RsXNFxWXPywLED1soBoo5pq8IlHFFn3awBblbRdcBMsVhYEdGK33yEJr96qcXhhhfvhXDw==","signature_status":"signed_v1","signed_at":"2026-08-06T01:35:42.858835Z","signed_message":"canonical_sha256_bytes"},"source_id":"2608.04496","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0b93fd1bb987ba2cb89c9d6867d680807585e41e343ebdfc8d42597b34becd25","sha256:bc5f5e9f57980c1db4a0c055fda157e47d451ef5bd72af59e2d8992a505c1ab2"],"state_sha256":"b78d295d52610609535cce97768b8e63c2bcb3878731f9dc53c02f3c29713770"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CR3VVSgaNiAy5N9nifazejx4GtAjVUhW7Rn2lkU3J4yLkYIIBZkBu1R67g0d8puskkCkJm03WuzK/Av1ze7wDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-24T02:09:05.884230Z","bundle_sha256":"b28f083e7825e5b70947e627a0940497b1d49d3d7707107923a6d2430aa5d884"}}