Pith Number

pith:USETYIEI

pith:2026:USETYIEIX4223A6DJFUDUUZU3Y

not attested not anchored not stored refs pending

OmniSIFT: Modality-Asymmetric Token Compression for Efficient Omni-modal Large Language Models

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

OmniSIFT reduces omni-modal token sequences to 25 percent of their length while matching or exceeding full-context accuracy on several tasks.

arxiv:2602.04804 v2 · 2026-02-04 · cs.CL

Open paper page JSON Open Graph Bundle Merged state Verified badge What is a Pith Number?

Add to your LaTeX paper

\usepackage{pith}
\pithnumber{USETYIEIX4223A6DJFUDUUZU3Y}

Prints a linked badge after your title and injects PDF metadata. Compiles on arXiv. Learn more · Embed verified badge

Record completeness

1 Bitcoin timestamp

2 Internet Archive

3 Author claim open · sign in to claim

4 Citations open

5 Replications open

✓ Portable graph bundle live · download bundle · merged state

The bundle contains the canonical record plus signed events. A mirror can host it anywhere and recompute the same current state with the deterministic merge algorithm.

Claims

C1strongest 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.

C2weakest 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.

C3one 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.

Cited by

7 papers in Pith

LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning

Stage-adaptive Token Selection for Efficient Omni-modal LLMs

OmniRefine: Alignment-Aware Cooperative Compression for Efficient Omnimodal Large Language Models

Keep What Audio Cannot Say: Context-Preserving Token Pruning for Omni-LLMs

HeadRouter: Dynamic Head-Weight Routing for Task-Adaptive Audio Token Pruning in Large Audio Language Models

Receipt and verification

First computed	2026-05-18T02:44:31.635686Z
Builder	pith-number-builder-2026-05-17-v1
Signature	Pith Ed25519 (`pith-v1-2026-05`) · public key
Schema	pith-number/v1.0

Canonical hash

a4893c2088bf35ad83c349683a5334de3db4316a4d314ba33de1b38737152192

Aliases

arxiv: 2602.04804 · arxiv_version: 2602.04804v2 · doi: 10.48550/arxiv.2602.04804 · pith_short_12: USETYIEIX422 · pith_short_16: USETYIEIX4223A6D · pith_short_8: USETYIEI

Agent API

Resolver JSON Graph JSON Events JSON Schema Signing key

Verify this Pith Number yourself

curl -sH 'Accept: application/ld+json' https://pith.science/pith/USETYIEIX4223A6DJFUDUUZU3Y \
  | jq -c '.canonical_record' \
  | python3 -c "import sys,json,hashlib; b=json.dumps(json.loads(sys.stdin.read()), sort_keys=True, separators=(',',':'), ensure_ascii=False).encode(); print(hashlib.sha256(b).hexdigest())"
# expect: a4893c2088bf35ad83c349683a5334de3db4316a4d314ba33de1b38737152192

Canonical record JSON

{
  "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
  }
}