{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:H4YFQVGEAAYQWZYAQBZCLCXZGB","short_pith_number":"pith:H4YFQVGE","canonical_record":{"source":{"id":"2504.09897","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-04-14T05:44:38Z","cross_cats_sorted":[],"title_canon_sha256":"93c1dceea8ad8406408178fab1ca967cf48537d1109a194db648bbf916779c11","abstract_canon_sha256":"efcceba76f11af1c5b5e0fb84282e050a2f6746768a33637be266930504f91a1"},"schema_version":"1.0"},"canonical_sha256":"3f305854c400310b67008072258af930539d7364219e376d50d53ae1937738a8","source":{"kind":"arxiv","id":"2504.09897","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.09897","created_at":"2026-07-05T11:04:48Z"},{"alias_kind":"arxiv_version","alias_value":"2504.09897v3","created_at":"2026-07-05T11:04:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.09897","created_at":"2026-07-05T11:04:48Z"},{"alias_kind":"pith_short_12","alias_value":"H4YFQVGEAAYQ","created_at":"2026-07-05T11:04:48Z"},{"alias_kind":"pith_short_16","alias_value":"H4YFQVGEAAYQWZYA","created_at":"2026-07-05T11:04:48Z"},{"alias_kind":"pith_short_8","alias_value":"H4YFQVGE","created_at":"2026-07-05T11:04:48Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:H4YFQVGEAAYQWZYAQBZCLCXZGB","target":"record","payload":{"canonical_record":{"source":{"id":"2504.09897","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-04-14T05:44:38Z","cross_cats_sorted":[],"title_canon_sha256":"93c1dceea8ad8406408178fab1ca967cf48537d1109a194db648bbf916779c11","abstract_canon_sha256":"efcceba76f11af1c5b5e0fb84282e050a2f6746768a33637be266930504f91a1"},"schema_version":"1.0"},"canonical_sha256":"3f305854c400310b67008072258af930539d7364219e376d50d53ae1937738a8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:04:48.978749Z","signature_b64":"ENQzwB+sLu0OLj9BSPKRAA9+b1bYEI+j9H+0Hqc9cKlt55+SCbt92EC8Zno84rtwTv1g5DdDiqFTztoHivdLAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3f305854c400310b67008072258af930539d7364219e376d50d53ae1937738a8","last_reissued_at":"2026-07-05T11:04:48.978196Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:04:48.978196Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2504.09897","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:04:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"qTO5FTD/CoqOvbRO9Gbbkn91Jempoz0nF5SweLVYAgjVNHZCAD3M1q9FSPvQcN8ZrpFhr9OchiQ/byjI0a3HDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-14T10:16:29.727271Z"},"content_sha256":"becc657d5e2b1df5ab3a7d57edbc0d645e1a29d5832e00eb37e16ced814e7404","schema_version":"1.0","event_id":"sha256:becc657d5e2b1df5ab3a7d57edbc0d645e1a29d5832e00eb37e16ced814e7404"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:H4YFQVGEAAYQWZYAQBZCLCXZGB","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"TAMP: Token-Adaptive Layerwise Pruning in Multimodal Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chanakya Ekbote, Jaewoo Lee, Keyang Xuan, Paul Pu Liang, Sandeep Polisetty, Yi R. Fung","submitted_at":"2025-04-14T05:44:38Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) have shown remarkable versatility in understanding diverse multimodal data and tasks. However, these capabilities come with an increased model scale. While post-training pruning reduces model size in unimodal models, its application to MLLMs often yields limited success. Our analysis discovers that conventional methods fail to account for the unique token attributes across layers and modalities inherent to MLLMs. Inspired by this observation, we propose TAMP, a simple yet effective pruning framework tailored for MLLMs, featuring two key components: (1) "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.09897","kind":"arxiv","version":3},"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/2504.09897/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":"2026-07-05T11:04:48Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"7sStnKImbdy6JwLYtB/8K4hyPffe0E52COI47yKufJOI0hv8YpfQAVuwbFaUu2AanLZ8zQ0XRKaMxwEfUYFUBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-14T10:16:29.727784Z"},"content_sha256":"b75fc48686e38942731caf3cd3da01d66c6815506c4cb212ad0e32e40421dde9","schema_version":"1.0","event_id":"sha256:b75fc48686e38942731caf3cd3da01d66c6815506c4cb212ad0e32e40421dde9"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/H4YFQVGEAAYQWZYAQBZCLCXZGB/bundle.json","state_url":"https://pith.science/pith/H4YFQVGEAAYQWZYAQBZCLCXZGB/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/H4YFQVGEAAYQWZYAQBZCLCXZGB/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":"2026-08-14T10:16:29Z","links":{"resolver":"https://pith.science/pith/H4YFQVGEAAYQWZYAQBZCLCXZGB","bundle":"https://pith.science/pith/H4YFQVGEAAYQWZYAQBZCLCXZGB/bundle.json","state":"https://pith.science/pith/H4YFQVGEAAYQWZYAQBZCLCXZGB/state.json","well_known_bundle":"https://pith.science/.well-known/pith/H4YFQVGEAAYQWZYAQBZCLCXZGB/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:H4YFQVGEAAYQWZYAQBZCLCXZGB","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":"efcceba76f11af1c5b5e0fb84282e050a2f6746768a33637be266930504f91a1","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-04-14T05:44:38Z","title_canon_sha256":"93c1dceea8ad8406408178fab1ca967cf48537d1109a194db648bbf916779c11"},"schema_version":"1.0","source":{"id":"2504.09897","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.09897","created_at":"2026-07-05T11:04:48Z"},{"alias_kind":"arxiv_version","alias_value":"2504.09897v3","created_at":"2026-07-05T11:04:48Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.09897","created_at":"2026-07-05T11:04:48Z"},{"alias_kind":"pith_short_12","alias_value":"H4YFQVGEAAYQ","created_at":"2026-07-05T11:04:48Z"},{"alias_kind":"pith_short_16","alias_value":"H4YFQVGEAAYQWZYA","created_at":"2026-07-05T11:04:48Z"},{"alias_kind":"pith_short_8","alias_value":"H4YFQVGE","created_at":"2026-07-05T11:04:48Z"}],"graph_snapshots":[{"event_id":"sha256:b75fc48686e38942731caf3cd3da01d66c6815506c4cb212ad0e32e40421dde9","target":"graph","created_at":"2026-07-05T11:04:48Z","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/2504.09897/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Multimodal Large Language Models (MLLMs) have shown remarkable versatility in understanding diverse multimodal data and tasks. However, these capabilities come with an increased model scale. While post-training pruning reduces model size in unimodal models, its application to MLLMs often yields limited success. Our analysis discovers that conventional methods fail to account for the unique token attributes across layers and modalities inherent to MLLMs. Inspired by this observation, we propose TAMP, a simple yet effective pruning framework tailored for MLLMs, featuring two key components: (1) ","authors_text":"Chanakya Ekbote, Jaewoo Lee, Keyang Xuan, Paul Pu Liang, Sandeep Polisetty, Yi R. Fung","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-04-14T05:44:38Z","title":"TAMP: Token-Adaptive Layerwise Pruning in Multimodal Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.09897","kind":"arxiv","version":3},"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:becc657d5e2b1df5ab3a7d57edbc0d645e1a29d5832e00eb37e16ced814e7404","target":"record","created_at":"2026-07-05T11:04:48Z","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":"efcceba76f11af1c5b5e0fb84282e050a2f6746768a33637be266930504f91a1","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-04-14T05:44:38Z","title_canon_sha256":"93c1dceea8ad8406408178fab1ca967cf48537d1109a194db648bbf916779c11"},"schema_version":"1.0","source":{"id":"2504.09897","kind":"arxiv","version":3}},"canonical_sha256":"3f305854c400310b67008072258af930539d7364219e376d50d53ae1937738a8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3f305854c400310b67008072258af930539d7364219e376d50d53ae1937738a8","first_computed_at":"2026-07-05T11:04:48.978196Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:04:48.978196Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ENQzwB+sLu0OLj9BSPKRAA9+b1bYEI+j9H+0Hqc9cKlt55+SCbt92EC8Zno84rtwTv1g5DdDiqFTztoHivdLAg==","signature_status":"signed_v1","signed_at":"2026-07-05T11:04:48.978749Z","signed_message":"canonical_sha256_bytes"},"source_id":"2504.09897","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:becc657d5e2b1df5ab3a7d57edbc0d645e1a29d5832e00eb37e16ced814e7404","sha256:b75fc48686e38942731caf3cd3da01d66c6815506c4cb212ad0e32e40421dde9"],"state_sha256":"f46a0135dbd1cf14647c04209c58be77b6dc602aba772ad57b0e0f39db2b3215"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1UG+e5h+i5m//7oozqWbFNX6iP1JQE0gTwU1ymdj5Tdbi5QPeks4W/fBjheHdWO8qUIZG2mrpLL8ulF490ciAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-14T10:16:29.732420Z","bundle_sha256":"9285ea2893786a7135d64a8038535dc564682594f8280cec3ebf38c65fb75c76"}}