{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:USETYIEIX4223A6DJFUDUUZU3Y","short_pith_number":"pith:USETYIEI","canonical_record":{"source":{"id":"2602.04804","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-04T17:51:05Z","cross_cats_sorted":[],"title_canon_sha256":"1b810817ebc3d1fb0e4f4d8c781cb53a299cbaf5deaa2f2ad26b2e117ea283df","abstract_canon_sha256":"752f62d099b3fa5a36b0b92e0ff005a1cc6de75a9612c0a1c333663bd761c5a8"},"schema_version":"1.0"},"canonical_sha256":"a4893c2088bf35ad83c349683a5334de3db4316a4d314ba33de1b38737152192","source":{"kind":"arxiv","id":"2602.04804","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.04804","created_at":"2026-05-18T02:44:31Z"},{"alias_kind":"arxiv_version","alias_value":"2602.04804v2","created_at":"2026-05-18T02:44:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.04804","created_at":"2026-05-18T02:44:31Z"},{"alias_kind":"pith_short_12","alias_value":"USETYIEIX422","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"USETYIEIX4223A6D","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"USETYIEI","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:USETYIEIX4223A6DJFUDUUZU3Y","target":"record","payload":{"canonical_record":{"source":{"id":"2602.04804","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-04T17:51:05Z","cross_cats_sorted":[],"title_canon_sha256":"1b810817ebc3d1fb0e4f4d8c781cb53a299cbaf5deaa2f2ad26b2e117ea283df","abstract_canon_sha256":"752f62d099b3fa5a36b0b92e0ff005a1cc6de75a9612c0a1c333663bd761c5a8"},"schema_version":"1.0"},"canonical_sha256":"a4893c2088bf35ad83c349683a5334de3db4316a4d314ba33de1b38737152192","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:31.636159Z","signature_b64":"SYbMikyR9d7frcjMd8cv9wZOff8uBKVv28aUIa9knAD3/9ac7KJ8HMRaduF6jFHZaXi7Kx4dlJlQgLpxgqEbAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a4893c2088bf35ad83c349683a5334de3db4316a4d314ba33de1b38737152192","last_reissued_at":"2026-05-18T02:44:31.635686Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:31.635686Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2602.04804","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":"2026-05-18T02:44:31Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cdSCjnfNd5xtYb+uXBjOWCxbzuFUA+cL4prYphscyA+A2irWxCX2tHjUwwjmsCBUGJSMHWvcUZnK9Uu7taYnCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-04T12:12:19.044510Z"},"content_sha256":"09f68efd90f7b30bde98914e575e5d9da28623ad4839254f82cd1fcb7504b1e5","schema_version":"1.0","event_id":"sha256:09f68efd90f7b30bde98914e575e5d9da28623ad4839254f82cd1fcb7504b1e5"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:USETYIEIX4223A6DJFUDUUZU3Y","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"OmniSIFT: Modality-Asymmetric Token Compression for Efficient Omni-modal Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Bohan Zeng, Bozhou Li, Jiaheng Liu, Junfei Wu, Jungang Li, Liang Wang, Pengfei Wan, Qiang Liu, Xinlong Chen, Xuyang Liu, Yang Shi, Yiyan Ji, Yuanxing Zhang, Yue Ding, Yushuo Guan","submitted_at":"2026-02-04T17:51:05Z","abstract_excerpt":"Omni-modal Large Language Models (Omni-LLMs) have demonstrated strong capabilities in audio-video understanding tasks. However, their reliance on long multimodal token sequences leads to substantial computational overhead. Despite this challenge, token compression methods designed for Omni-LLMs remain limited. To bridge this gap, we propose OmniSIFT (Omni-modal Spatio-temporal Informed Fine-grained Token compression), a modality-asymmetric token compression framework tailored for Omni-LLMs. Specifically, OmniSIFT adopts a two-stage compression strategy: (i) a spatio-temporal video pruning modu"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"With merely 25% of the original token context, OmniSIFT consistently outperforms all compression baselines and even surpasses the performance of the full-token model on several tasks.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the spatio-temporal pruning and vision-guided audio selection preserve all task-critical information across diverse video-audio content without introducing systematic bias.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"OmniSIFT delivers modality-asymmetric token compression for omni-LLMs via spatio-temporal video pruning and vision-guided audio selection, outperforming baselines at 25% token count.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f84376d72f1a270cb1af84fc693fa689699909ab1be784bb1ed8f4afd857eea7"},"source":{"id":"2602.04804","kind":"arxiv","version":2},"verdict":{"id":"1c2b82ee-18a6-486d-8527-6019016f34fe","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T07:21:07.168044Z","strongest_claim":"With merely 25% of the original token context, OmniSIFT consistently outperforms all compression baselines and even surpasses the performance of the full-token model on several tasks.","one_line_summary":"OmniSIFT delivers modality-asymmetric token compression for omni-LLMs via spatio-temporal video pruning and vision-guided audio selection, outperforming baselines at 25% token count.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the spatio-temporal pruning and vision-guided audio selection preserve all task-critical information across diverse video-audio content without introducing systematic bias.","pith_extraction_headline":"OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks."},"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":"1c2b82ee-18a6-486d-8527-6019016f34fe"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:31Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"WIUDkj9V2Uh9uOxlxxr8el/MMsu7RN95xrEYQcuHsClIt75TpVuWyn/C3PkEkNSkWMkQDzpsQ8M4EjUz38pBAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-04T12:12:19.044983Z"},"content_sha256":"151641de1af216e52259bf0dd0129b35f581cfcbd0d45ea70179294edccf8eb5","schema_version":"1.0","event_id":"sha256:151641de1af216e52259bf0dd0129b35f581cfcbd0d45ea70179294edccf8eb5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y/bundle.json","state_url":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/USETYIEIX4223A6DJFUDUUZU3Y/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":"2026-06-04T12:12:19Z","links":{"resolver":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y","bundle":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y/bundle.json","state":"https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y/state.json","well_known_bundle":"https://pith.science/.well-known/pith/USETYIEIX4223A6DJFUDUUZU3Y/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:USETYIEIX4223A6DJFUDUUZU3Y","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":"752f62d099b3fa5a36b0b92e0ff005a1cc6de75a9612c0a1c333663bd761c5a8","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-04T17:51:05Z","title_canon_sha256":"1b810817ebc3d1fb0e4f4d8c781cb53a299cbaf5deaa2f2ad26b2e117ea283df"},"schema_version":"1.0","source":{"id":"2602.04804","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.04804","created_at":"2026-05-18T02:44:31Z"},{"alias_kind":"arxiv_version","alias_value":"2602.04804v2","created_at":"2026-05-18T02:44:31Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.04804","created_at":"2026-05-18T02:44:31Z"},{"alias_kind":"pith_short_12","alias_value":"USETYIEIX422","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"USETYIEIX4223A6D","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"USETYIEI","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:151641de1af216e52259bf0dd0129b35f581cfcbd0d45ea70179294edccf8eb5","target":"graph","created_at":"2026-05-18T02:44:31Z","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":"With merely 25% of the original token context, OmniSIFT consistently outperforms all compression baselines and even surpasses the performance of the full-token model on several tasks."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the spatio-temporal pruning and vision-guided audio selection preserve all task-critical information across diverse video-audio content without introducing systematic bias."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"OmniSIFT delivers modality-asymmetric token compression for omni-LLMs via spatio-temporal video pruning and vision-guided audio selection, outperforming baselines at 25% token count."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks."}],"snapshot_sha256":"f84376d72f1a270cb1af84fc693fa689699909ab1be784bb1ed8f4afd857eea7"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Omni-modal Large Language Models (Omni-LLMs) have demonstrated strong capabilities in audio-video understanding tasks. However, their reliance on long multimodal token sequences leads to substantial computational overhead. Despite this challenge, token compression methods designed for Omni-LLMs remain limited. To bridge this gap, we propose OmniSIFT (Omni-modal Spatio-temporal Informed Fine-grained Token compression), a modality-asymmetric token compression framework tailored for Omni-LLMs. Specifically, OmniSIFT adopts a two-stage compression strategy: (i) a spatio-temporal video pruning modu","authors_text":"Bohan Zeng, Bozhou Li, Jiaheng Liu, Junfei Wu, Jungang Li, Liang Wang, Pengfei Wan, Qiang Liu, Xinlong Chen, Xuyang Liu, Yang Shi, Yiyan Ji, Yuanxing Zhang, Yue Ding, Yushuo Guan","cross_cats":[],"headline":"OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-04T17:51:05Z","title":"OmniSIFT: Modality-Asymmetric Token Compression for Efficient Omni-modal Large Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.04804","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-16T07:21:07.168044Z","id":"1c2b82ee-18a6-486d-8527-6019016f34fe","model_set":{"reader":"grok-4.3"},"one_line_summary":"OmniSIFT delivers modality-asymmetric token compression for omni-LLMs via spatio-temporal video pruning and vision-guided audio selection, outperforming baselines at 25% token count.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks.","strongest_claim":"With merely 25% of the original token context, OmniSIFT consistently outperforms all compression baselines and even surpasses the performance of the full-token model on several tasks.","weakest_assumption":"That the spatio-temporal pruning and vision-guided audio selection preserve all task-critical information across diverse video-audio content without introducing systematic bias."}},"verdict_id":"1c2b82ee-18a6-486d-8527-6019016f34fe"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:09f68efd90f7b30bde98914e575e5d9da28623ad4839254f82cd1fcb7504b1e5","target":"record","created_at":"2026-05-18T02:44:31Z","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":"752f62d099b3fa5a36b0b92e0ff005a1cc6de75a9612c0a1c333663bd761c5a8","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-02-04T17:51:05Z","title_canon_sha256":"1b810817ebc3d1fb0e4f4d8c781cb53a299cbaf5deaa2f2ad26b2e117ea283df"},"schema_version":"1.0","source":{"id":"2602.04804","kind":"arxiv","version":2}},"canonical_sha256":"a4893c2088bf35ad83c349683a5334de3db4316a4d314ba33de1b38737152192","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a4893c2088bf35ad83c349683a5334de3db4316a4d314ba33de1b38737152192","first_computed_at":"2026-05-18T02:44:31.635686Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:31.635686Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"SYbMikyR9d7frcjMd8cv9wZOff8uBKVv28aUIa9knAD3/9ac7KJ8HMRaduF6jFHZaXi7Kx4dlJlQgLpxgqEbAw==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:31.636159Z","signed_message":"canonical_sha256_bytes"},"source_id":"2602.04804","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:09f68efd90f7b30bde98914e575e5d9da28623ad4839254f82cd1fcb7504b1e5","sha256:151641de1af216e52259bf0dd0129b35f581cfcbd0d45ea70179294edccf8eb5"],"state_sha256":"ca5270af36721d3ba30e2e14dbe81d7b61b9366e5d0271b717d462e3c6c9d440"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"1OSlGEesQCOFeVV72nsql59OW6VLyE7fVv6aqM1FkkbbzPDZNMHOv7wDyvKt7RdPa25NklLcjnuWDBP+gCX0Ag==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-04T12:12:19.047196Z","bundle_sha256":"5d6aabc7ef2b22d65e737bcd1202e1caf6f128740f52aa5c0129f8584e2cdb7e"}}