{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:FRMVISTZYKVKUDNC2WWYZR6KUZ","short_pith_number":"pith:FRMVISTZ","schema_version":"1.0","canonical_sha256":"2c59544a79c2aaaa0da2d5ad8cc7caa65612c53f856a87bf87209e79b8088540","source":{"kind":"arxiv","id":"2606.09508","version":1},"attestation_state":"computed","paper":{"title":"From Rigid to Dynamic: Entropy-Guided Adaptive Inference for Long-Context LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Chen Jason Zhang, Fei Teng, Haoyang Li, Lei Chen, Qingfa Xiao, Qing Li, Zhanchao Xu","submitted_at":"2026-06-08T14:02:18Z","abstract_excerpt":"Existing sparse attention and KV cache compression methods for long-context LLM inference typically apply fixed sparsity patterns or uniform budgets across all attention heads, overlooking the substantial variation in attention behavior among heads and contexts. We observe two distinct entropy patterns among attention heads: Rigid Heads, whose entropy stays near zero across input segments, and Dynamic Heads, whose entropy fluctuates significantly. Crucially, the distribution of these types is context-dependent and cannot be predetermined offline. We therefore propose EntropyInfer, a training-f"},"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":"2606.09508","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-08T14:02:18Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"c9f573a469f05c4b65d86354de94c52b7d26ffe80fe168f488189f316b44f5de","abstract_canon_sha256":"023f1154e6526b9b3e7468bc03d79e8f7cbae81119792771fef6ea7bda49ded8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-09T02:08:52.576743Z","signature_b64":"CgB0pF/FyVf3NO+BOSP1n4mc/U6oLKA/Kq4HI50rXEVFb+5gTx0rFTmE+0PdVOxuYDzeOFJdUz7eM5gYYFBIBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2c59544a79c2aaaa0da2d5ad8cc7caa65612c53f856a87bf87209e79b8088540","last_reissued_at":"2026-06-09T02:08:52.575893Z","signature_status":"signed_v1","first_computed_at":"2026-06-09T02:08:52.575893Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From Rigid to Dynamic: Entropy-Guided Adaptive Inference for Long-Context LLMs","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.AI","authors_text":"Chen Jason Zhang, Fei Teng, Haoyang Li, Lei Chen, Qingfa Xiao, Qing Li, Zhanchao Xu","submitted_at":"2026-06-08T14:02:18Z","abstract_excerpt":"Existing sparse attention and KV cache compression methods for long-context LLM inference typically apply fixed sparsity patterns or uniform budgets across all attention heads, overlooking the substantial variation in attention behavior among heads and contexts. We observe two distinct entropy patterns among attention heads: Rigid Heads, whose entropy stays near zero across input segments, and Dynamic Heads, whose entropy fluctuates significantly. Crucially, the distribution of these types is context-dependent and cannot be predetermined offline. We therefore propose EntropyInfer, a training-f"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.09508","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/2606.09508/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":"2606.09508","created_at":"2026-06-09T02:08:52.576040+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.09508v1","created_at":"2026-06-09T02:08:52.576040+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.09508","created_at":"2026-06-09T02:08:52.576040+00:00"},{"alias_kind":"pith_short_12","alias_value":"FRMVISTZYKVK","created_at":"2026-06-09T02:08:52.576040+00:00"},{"alias_kind":"pith_short_16","alias_value":"FRMVISTZYKVKUDNC","created_at":"2026-06-09T02:08:52.576040+00:00"},{"alias_kind":"pith_short_8","alias_value":"FRMVISTZ","created_at":"2026-06-09T02:08:52.576040+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/FRMVISTZYKVKUDNC2WWYZR6KUZ","json":"https://pith.science/pith/FRMVISTZYKVKUDNC2WWYZR6KUZ.json","graph_json":"https://pith.science/api/pith-number/FRMVISTZYKVKUDNC2WWYZR6KUZ/graph.json","events_json":"https://pith.science/api/pith-number/FRMVISTZYKVKUDNC2WWYZR6KUZ/events.json","paper":"https://pith.science/paper/FRMVISTZ"},"agent_actions":{"view_html":"https://pith.science/pith/FRMVISTZYKVKUDNC2WWYZR6KUZ","download_json":"https://pith.science/pith/FRMVISTZYKVKUDNC2WWYZR6KUZ.json","view_paper":"https://pith.science/paper/FRMVISTZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.09508&json=true","fetch_graph":"https://pith.science/api/pith-number/FRMVISTZYKVKUDNC2WWYZR6KUZ/graph.json","fetch_events":"https://pith.science/api/pith-number/FRMVISTZYKVKUDNC2WWYZR6KUZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FRMVISTZYKVKUDNC2WWYZR6KUZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FRMVISTZYKVKUDNC2WWYZR6KUZ/action/storage_attestation","attest_author":"https://pith.science/pith/FRMVISTZYKVKUDNC2WWYZR6KUZ/action/author_attestation","sign_citation":"https://pith.science/pith/FRMVISTZYKVKUDNC2WWYZR6KUZ/action/citation_signature","submit_replication":"https://pith.science/pith/FRMVISTZYKVKUDNC2WWYZR6KUZ/action/replication_record"}},"created_at":"2026-06-09T02:08:52.576040+00:00","updated_at":"2026-06-09T02:08:52.576040+00:00"}