{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:S2XR4CRX3PY536M7SK2UTUZPAU","short_pith_number":"pith:S2XR4CRX","schema_version":"1.0","canonical_sha256":"96af1e0a37dbf1ddf99f92b549d32f051dc7c0d5ad64baffb2c9b95b1bcd08c8","source":{"kind":"arxiv","id":"2606.31982","version":1},"attestation_state":"computed","paper":{"title":"ERA: Entropy-Guided Visual Token Pruning with Rectified Attention for Efficient MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Haiwen Diao, Huchuan Lu, Lei Zhang, Mu Qiao, Pingping Zhang, Xindong Zhang, Yuhao Wang, Yunzhi Zhuge","submitted_at":"2026-06-30T17:20:29Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) incur prohibitive inference costs due to long visual token sequences. Training-free visual token reduction provides an efficient solution. However, existing methods distort attention distributions, giving rise to a phenomenon we term Attention Logit Collapse. To address this issue, we propose ERA, an Entropy-guided visual token pruning framework with Rectified Attention for efficient MLLMs. Specifically, ERA comprises three crucial components: Dual-view Entropy Pruning (DEP), Bias-aware Token Recycling (BTR), and Logit-preserving Attention Rectification"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.31982","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-06-30T17:20:29Z","cross_cats_sorted":[],"title_canon_sha256":"8e538d6936f55e1257a07170361edd898d13b4e3647fe25a20a6c7c4e4cc1f9c","abstract_canon_sha256":"330655314c3c0e167d8627c26d855622fe43e81d8795b39be73ecfcba5270365"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-01T01:18:27.782214Z","signature_b64":"tG2hvvFqDDdWRr88JtM8EcXvm/ebKjXpssjmVOxmXbiDKHI+zhcu8nwiUywNBPum4rcxCfC+EB5hpzL8+RfkBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"96af1e0a37dbf1ddf99f92b549d32f051dc7c0d5ad64baffb2c9b95b1bcd08c8","last_reissued_at":"2026-07-01T01:18:27.781672Z","signature_status":"signed_v1","first_computed_at":"2026-07-01T01:18:27.781672Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ERA: Entropy-Guided Visual Token Pruning with Rectified Attention for Efficient MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Haiwen Diao, Huchuan Lu, Lei Zhang, Mu Qiao, Pingping Zhang, Xindong Zhang, Yuhao Wang, Yunzhi Zhuge","submitted_at":"2026-06-30T17:20:29Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) incur prohibitive inference costs due to long visual token sequences. Training-free visual token reduction provides an efficient solution. However, existing methods distort attention distributions, giving rise to a phenomenon we term Attention Logit Collapse. To address this issue, we propose ERA, an Entropy-guided visual token pruning framework with Rectified Attention for efficient MLLMs. Specifically, ERA comprises three crucial components: Dual-view Entropy Pruning (DEP), Bias-aware Token Recycling (BTR), and Logit-preserving Attention Rectification"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.31982","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/2606.31982/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.31982","created_at":"2026-07-01T01:18:27.781747+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.31982v1","created_at":"2026-07-01T01:18:27.781747+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.31982","created_at":"2026-07-01T01:18:27.781747+00:00"},{"alias_kind":"pith_short_12","alias_value":"S2XR4CRX3PY5","created_at":"2026-07-01T01:18:27.781747+00:00"},{"alias_kind":"pith_short_16","alias_value":"S2XR4CRX3PY536M7","created_at":"2026-07-01T01:18:27.781747+00:00"},{"alias_kind":"pith_short_8","alias_value":"S2XR4CRX","created_at":"2026-07-01T01:18:27.781747+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/S2XR4CRX3PY536M7SK2UTUZPAU","json":"https://pith.science/pith/S2XR4CRX3PY536M7SK2UTUZPAU.json","graph_json":"https://pith.science/api/pith-number/S2XR4CRX3PY536M7SK2UTUZPAU/graph.json","events_json":"https://pith.science/api/pith-number/S2XR4CRX3PY536M7SK2UTUZPAU/events.json","paper":"https://pith.science/paper/S2XR4CRX"},"agent_actions":{"view_html":"https://pith.science/pith/S2XR4CRX3PY536M7SK2UTUZPAU","download_json":"https://pith.science/pith/S2XR4CRX3PY536M7SK2UTUZPAU.json","view_paper":"https://pith.science/paper/S2XR4CRX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.31982&json=true","fetch_graph":"https://pith.science/api/pith-number/S2XR4CRX3PY536M7SK2UTUZPAU/graph.json","fetch_events":"https://pith.science/api/pith-number/S2XR4CRX3PY536M7SK2UTUZPAU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/S2XR4CRX3PY536M7SK2UTUZPAU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/S2XR4CRX3PY536M7SK2UTUZPAU/action/storage_attestation","attest_author":"https://pith.science/pith/S2XR4CRX3PY536M7SK2UTUZPAU/action/author_attestation","sign_citation":"https://pith.science/pith/S2XR4CRX3PY536M7SK2UTUZPAU/action/citation_signature","submit_replication":"https://pith.science/pith/S2XR4CRX3PY536M7SK2UTUZPAU/action/replication_record"}},"created_at":"2026-07-01T01:18:27.781747+00:00","updated_at":"2026-07-01T01:18:27.781747+00:00"}