{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:EC2YM236SUOV5K37B7OX6EY6KI","short_pith_number":"pith:EC2YM236","schema_version":"1.0","canonical_sha256":"20b5866b7e951d5eab7f0fdd7f131e5225f0d0ab5b0b46e79b7a11f20a650867","source":{"kind":"arxiv","id":"2608.01979","version":1},"attestation_state":"computed","paper":{"title":"ET-Prune: Evidence-Aware Dynamic Budgeting for Visual Token Pruning in Text-Rich MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Junxian Li, Kai Liu, Linghe Kong, Shaoqiu Zhang, Xiao Xiao, Yulun Zhang, Zizhong Ding","submitted_at":"2026-08-03T09:42:34Z","abstract_excerpt":"Visual token pruning reduces the inference cost of multimodal large language models, but a fixed token ratio is poorly matched to text-rich inputs. In OCR-centric tasks, decisive evidence can be a small number, label, or field whose relevance is specified by the question; indiscriminate pruning can erase that evidence while retaining visually salient but irrelevant regions. We present ET-Prune, a training-free framework that casts pruning as evidence allocation. It derives question-conditioned evidence from a decoder-side partial query-key block, safeguards text-like spatial regions, and conve"},"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":"2608.01979","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-08-03T09:42:34Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"89e5fd6c196bea880af5d4abe61b5f8b181fbd65feb2b77e6dfe6e1b9738aa68","abstract_canon_sha256":"cce8815979af80089fe9e01a292d7542895b79a2fde395a73f9cd74500adba99"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-04T02:10:25.951837Z","signature_b64":"ejfhMkoFb9zUQQVGdfPGi4SPLkz17wtPY1k0mdl96WmNdPR/HdSH89o8TlPPpBdoaY0av3rty3Cj9wGmIpNcDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"20b5866b7e951d5eab7f0fdd7f131e5225f0d0ab5b0b46e79b7a11f20a650867","last_reissued_at":"2026-08-04T02:10:25.950046Z","signature_status":"signed_v1","first_computed_at":"2026-08-04T02:10:25.950046Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ET-Prune: Evidence-Aware Dynamic Budgeting for Visual Token Pruning in Text-Rich MLLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Junxian Li, Kai Liu, Linghe Kong, Shaoqiu Zhang, Xiao Xiao, Yulun Zhang, Zizhong Ding","submitted_at":"2026-08-03T09:42:34Z","abstract_excerpt":"Visual token pruning reduces the inference cost of multimodal large language models, but a fixed token ratio is poorly matched to text-rich inputs. In OCR-centric tasks, decisive evidence can be a small number, label, or field whose relevance is specified by the question; indiscriminate pruning can erase that evidence while retaining visually salient but irrelevant regions. We present ET-Prune, a training-free framework that casts pruning as evidence allocation. It derives question-conditioned evidence from a decoder-side partial query-key block, safeguards text-like spatial regions, and conve"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.01979","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/2608.01979/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":"2608.01979","created_at":"2026-08-04T02:10:25.951743+00:00"},{"alias_kind":"arxiv_version","alias_value":"2608.01979v1","created_at":"2026-08-04T02:10:25.951743+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.01979","created_at":"2026-08-04T02:10:25.951743+00:00"},{"alias_kind":"pith_short_12","alias_value":"EC2YM236SUOV","created_at":"2026-08-04T02:10:25.951743+00:00"},{"alias_kind":"pith_short_16","alias_value":"EC2YM236SUOV5K37","created_at":"2026-08-04T02:10:25.951743+00:00"},{"alias_kind":"pith_short_8","alias_value":"EC2YM236","created_at":"2026-08-04T02:10:25.951743+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/EC2YM236SUOV5K37B7OX6EY6KI","json":"https://pith.science/pith/EC2YM236SUOV5K37B7OX6EY6KI.json","graph_json":"https://pith.science/api/pith-number/EC2YM236SUOV5K37B7OX6EY6KI/graph.json","events_json":"https://pith.science/api/pith-number/EC2YM236SUOV5K37B7OX6EY6KI/events.json","paper":"https://pith.science/paper/EC2YM236"},"agent_actions":{"view_html":"https://pith.science/pith/EC2YM236SUOV5K37B7OX6EY6KI","download_json":"https://pith.science/pith/EC2YM236SUOV5K37B7OX6EY6KI.json","view_paper":"https://pith.science/paper/EC2YM236","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2608.01979&json=true","fetch_graph":"https://pith.science/api/pith-number/EC2YM236SUOV5K37B7OX6EY6KI/graph.json","fetch_events":"https://pith.science/api/pith-number/EC2YM236SUOV5K37B7OX6EY6KI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EC2YM236SUOV5K37B7OX6EY6KI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EC2YM236SUOV5K37B7OX6EY6KI/action/storage_attestation","attest_author":"https://pith.science/pith/EC2YM236SUOV5K37B7OX6EY6KI/action/author_attestation","sign_citation":"https://pith.science/pith/EC2YM236SUOV5K37B7OX6EY6KI/action/citation_signature","submit_replication":"https://pith.science/pith/EC2YM236SUOV5K37B7OX6EY6KI/action/replication_record"}},"created_at":"2026-08-04T02:10:25.951743+00:00","updated_at":"2026-08-04T02:10:25.951743+00:00"}