{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:V5T7MQS44HEDIYTU34VSSMKCBS","short_pith_number":"pith:V5T7MQS4","canonical_record":{"source":{"id":"2412.03248","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-12-04T11:47:57Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"90fe4c5aa177aad9c6559f2296d8a8ac2bc8b2f71e3865263dc56bf4fa306758","abstract_canon_sha256":"ed24e38918957e46d2ac55346fd7f716dbfb48c447906314976e41a21a033abd"},"schema_version":"1.0"},"canonical_sha256":"af67f6425ce1c8346274df2b2931420cb5a7ea509550ef10af548afc84adf1f8","source":{"kind":"arxiv","id":"2412.03248","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2412.03248","created_at":"2026-07-05T11:44:43Z"},{"alias_kind":"arxiv_version","alias_value":"2412.03248v2","created_at":"2026-07-05T11:44:43Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.03248","created_at":"2026-07-05T11:44:43Z"},{"alias_kind":"pith_short_12","alias_value":"V5T7MQS44HED","created_at":"2026-07-05T11:44:43Z"},{"alias_kind":"pith_short_16","alias_value":"V5T7MQS44HEDIYTU","created_at":"2026-07-05T11:44:43Z"},{"alias_kind":"pith_short_8","alias_value":"V5T7MQS4","created_at":"2026-07-05T11:44:43Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:V5T7MQS44HEDIYTU34VSSMKCBS","target":"record","payload":{"canonical_record":{"source":{"id":"2412.03248","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-12-04T11:47:57Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"90fe4c5aa177aad9c6559f2296d8a8ac2bc8b2f71e3865263dc56bf4fa306758","abstract_canon_sha256":"ed24e38918957e46d2ac55346fd7f716dbfb48c447906314976e41a21a033abd"},"schema_version":"1.0"},"canonical_sha256":"af67f6425ce1c8346274df2b2931420cb5a7ea509550ef10af548afc84adf1f8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:44:43.299778Z","signature_b64":"iuSvk+RYDTWRL2SFBwZeTbfSbvyLzXVxWFe7q1QDsJaHMGltPqQRvOOYzA5u0MXdpWrTDOmoRZ6Me0M6SeaZAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"af67f6425ce1c8346274df2b2931420cb5a7ea509550ef10af548afc84adf1f8","last_reissued_at":"2026-07-05T11:44:43.299275Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:44:43.299275Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2412.03248","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":"2026-07-05T11:44:43Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5bFN+ZwknrB6P5dszcGJoHosr1ifLEfbXaChjCE5aymFGgDb8KkDK4Z5k37HGcXQKazY67eWwuFzIi39fwr5Ag==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T20:58:51.376540Z"},"content_sha256":"b54c9d34fa05aa02413ba1a068a82e9c9e632379a8d19f15bb13171228f7ee38","schema_version":"1.0","event_id":"sha256:b54c9d34fa05aa02413ba1a068a82e9c9e632379a8d19f15bb13171228f7ee38"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:V5T7MQS44HEDIYTU34VSSMKCBS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"AIM: Adaptive Inference of Multi-Modal LLMs via Token Merging and Pruning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.CV","authors_text":"Liwei Wang, Yin Li, Yiwu Zhong, Zhuoming Liu","submitted_at":"2024-12-04T11:47:57Z","abstract_excerpt":"Large language models (LLMs) have enabled the creation of multi-modal LLMs that exhibit strong comprehension of visual data such as images and videos. However, these models usually rely on extensive visual tokens from visual encoders, leading to high computational demands, which limits their applicability in resource-constrained environments and for long-context tasks. In this work, we propose a training-free adaptive inference method for multi-modal LLMs that can accommodate a broad range of efficiency requirements with a minimum performance drop. Our method consists of a) iterative token mer"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.03248","kind":"arxiv","version":2},"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/2412.03248/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":"2026-07-05T11:44:43Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"q14SIYgi/0hQrwHyRc6J6Kg9QgsAh/QI2OHGe6hPjdLBwNo9loF3VrK+4m+hFjr/C1o4PKzbKWrF02gMsb5pCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T20:58:51.377073Z"},"content_sha256":"8e29e43a0b7c154a2b568a10c4017dd0561589eae8046435f771c6557503a181","schema_version":"1.0","event_id":"sha256:8e29e43a0b7c154a2b568a10c4017dd0561589eae8046435f771c6557503a181"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/V5T7MQS44HEDIYTU34VSSMKCBS/bundle.json","state_url":"https://pith.science/pith/V5T7MQS44HEDIYTU34VSSMKCBS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/V5T7MQS44HEDIYTU34VSSMKCBS/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":"2026-08-08T20:58:51Z","links":{"resolver":"https://pith.science/pith/V5T7MQS44HEDIYTU34VSSMKCBS","bundle":"https://pith.science/pith/V5T7MQS44HEDIYTU34VSSMKCBS/bundle.json","state":"https://pith.science/pith/V5T7MQS44HEDIYTU34VSSMKCBS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/V5T7MQS44HEDIYTU34VSSMKCBS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:V5T7MQS44HEDIYTU34VSSMKCBS","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":"ed24e38918957e46d2ac55346fd7f716dbfb48c447906314976e41a21a033abd","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-12-04T11:47:57Z","title_canon_sha256":"90fe4c5aa177aad9c6559f2296d8a8ac2bc8b2f71e3865263dc56bf4fa306758"},"schema_version":"1.0","source":{"id":"2412.03248","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2412.03248","created_at":"2026-07-05T11:44:43Z"},{"alias_kind":"arxiv_version","alias_value":"2412.03248v2","created_at":"2026-07-05T11:44:43Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.03248","created_at":"2026-07-05T11:44:43Z"},{"alias_kind":"pith_short_12","alias_value":"V5T7MQS44HED","created_at":"2026-07-05T11:44:43Z"},{"alias_kind":"pith_short_16","alias_value":"V5T7MQS44HEDIYTU","created_at":"2026-07-05T11:44:43Z"},{"alias_kind":"pith_short_8","alias_value":"V5T7MQS4","created_at":"2026-07-05T11:44:43Z"}],"graph_snapshots":[{"event_id":"sha256:8e29e43a0b7c154a2b568a10c4017dd0561589eae8046435f771c6557503a181","target":"graph","created_at":"2026-07-05T11:44:43Z","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/2412.03248/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) have enabled the creation of multi-modal LLMs that exhibit strong comprehension of visual data such as images and videos. However, these models usually rely on extensive visual tokens from visual encoders, leading to high computational demands, which limits their applicability in resource-constrained environments and for long-context tasks. In this work, we propose a training-free adaptive inference method for multi-modal LLMs that can accommodate a broad range of efficiency requirements with a minimum performance drop. Our method consists of a) iterative token mer","authors_text":"Liwei Wang, Yin Li, Yiwu Zhong, Zhuoming Liu","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-12-04T11:47:57Z","title":"AIM: Adaptive Inference of Multi-Modal LLMs via Token Merging and Pruning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.03248","kind":"arxiv","version":2},"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:b54c9d34fa05aa02413ba1a068a82e9c9e632379a8d19f15bb13171228f7ee38","target":"record","created_at":"2026-07-05T11:44:43Z","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":"ed24e38918957e46d2ac55346fd7f716dbfb48c447906314976e41a21a033abd","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-12-04T11:47:57Z","title_canon_sha256":"90fe4c5aa177aad9c6559f2296d8a8ac2bc8b2f71e3865263dc56bf4fa306758"},"schema_version":"1.0","source":{"id":"2412.03248","kind":"arxiv","version":2}},"canonical_sha256":"af67f6425ce1c8346274df2b2931420cb5a7ea509550ef10af548afc84adf1f8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"af67f6425ce1c8346274df2b2931420cb5a7ea509550ef10af548afc84adf1f8","first_computed_at":"2026-07-05T11:44:43.299275Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:44:43.299275Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"iuSvk+RYDTWRL2SFBwZeTbfSbvyLzXVxWFe7q1QDsJaHMGltPqQRvOOYzA5u0MXdpWrTDOmoRZ6Me0M6SeaZAA==","signature_status":"signed_v1","signed_at":"2026-07-05T11:44:43.299778Z","signed_message":"canonical_sha256_bytes"},"source_id":"2412.03248","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b54c9d34fa05aa02413ba1a068a82e9c9e632379a8d19f15bb13171228f7ee38","sha256:8e29e43a0b7c154a2b568a10c4017dd0561589eae8046435f771c6557503a181"],"state_sha256":"e8bfc45df0ff2a02abea73e3c9660a4c53bdd1af615d925b484012eb0c40f041"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"AfatHiT3OivCA7bsueOMwUX4tIZ4oTNHf14wzF38MxsUYCB7+7qCn1XJsSFcGbSYkbFyuFEs9NYlxW2vgsLcBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T20:58:51.382793Z","bundle_sha256":"57878b22e8fab5ae432d2b00829a9c2452737183c1d2b9d7c431e0fe471245cb"}}