{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:CCYRHP5GUMOWSPXXZN4ZQKQPTW","short_pith_number":"pith:CCYRHP5G","canonical_record":{"source":{"id":"2310.17157","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-10-26T05:01:09Z","cross_cats_sorted":[],"title_canon_sha256":"a2373d8298a6249c1540fe27bca747e5ad6d86b0090ca87607785be58ddcdac1","abstract_canon_sha256":"8f2c97e7e1e8daac2c8daec1e2dacdfeff9af44b31b93bac054bca6b39575a05"},"schema_version":"1.0"},"canonical_sha256":"10b113bfa6a31d693ef7cb79982a0f9d86b27afe8ab7b2aac828535a68a5f554","source":{"kind":"arxiv","id":"2310.17157","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.17157","created_at":"2026-07-05T07:05:20Z"},{"alias_kind":"arxiv_version","alias_value":"2310.17157v1","created_at":"2026-07-05T07:05:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.17157","created_at":"2026-07-05T07:05:20Z"},{"alias_kind":"pith_short_12","alias_value":"CCYRHP5GUMOW","created_at":"2026-07-05T07:05:20Z"},{"alias_kind":"pith_short_16","alias_value":"CCYRHP5GUMOWSPXX","created_at":"2026-07-05T07:05:20Z"},{"alias_kind":"pith_short_8","alias_value":"CCYRHP5G","created_at":"2026-07-05T07:05:20Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:CCYRHP5GUMOWSPXXZN4ZQKQPTW","target":"record","payload":{"canonical_record":{"source":{"id":"2310.17157","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-10-26T05:01:09Z","cross_cats_sorted":[],"title_canon_sha256":"a2373d8298a6249c1540fe27bca747e5ad6d86b0090ca87607785be58ddcdac1","abstract_canon_sha256":"8f2c97e7e1e8daac2c8daec1e2dacdfeff9af44b31b93bac054bca6b39575a05"},"schema_version":"1.0"},"canonical_sha256":"10b113bfa6a31d693ef7cb79982a0f9d86b27afe8ab7b2aac828535a68a5f554","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:05:20.761187Z","signature_b64":"Tsm1ehF9q1E3BlHgfNFekmUmDOzfLJXS5S/fUf9Pr3n64O/prM3B4+rcz614L3lPNChliR78ptHUz886luHcCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"10b113bfa6a31d693ef7cb79982a0f9d86b27afe8ab7b2aac828535a68a5f554","last_reissued_at":"2026-07-05T07:05:20.760782Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:05:20.760782Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2310.17157","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:05:20Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"y4sqD/QL2D9ghdtKEn0TYyeLnUXyz3sR6Kh9omIgIgFn7BIQt7ZKJga5/1DrPELkrU0Lfzum1pnzLyWqXp+6Dg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-14T23:14:50.761112Z"},"content_sha256":"4e4c540998288a4fee14bed27f9531c087f4176c24b34cd96c03b532a4718a2a","schema_version":"1.0","event_id":"sha256:4e4c540998288a4fee14bed27f9531c087f4176c24b34cd96c03b532a4718a2a"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:CCYRHP5GUMOWSPXXZN4ZQKQPTW","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Deja Vu: Contextual Sparsity for Efficient LLMs at Inference Time","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anshumali Shrivastava, Beidi Chen, Binhang Yuan, Ce Zhang, Christopher Re, Jue Wang, Tianyi Zhou, Tri Dao, Yuandong Tian, Zhao Song, Zichang Liu","submitted_at":"2023-10-26T05:01:09Z","abstract_excerpt":"Large language models (LLMs) with hundreds of billions of parameters have sparked a new wave of exciting AI applications. However, they are computationally expensive at inference time. Sparsity is a natural approach to reduce this cost, but existing methods either require costly retraining, have to forgo LLM's in-context learning ability, or do not yield wall-clock time speedup on modern hardware. We hypothesize that contextual sparsity, which are small, input-dependent sets of attention heads and MLP parameters that yield approximately the same output as the dense model for a given input, can"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.17157","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/2310.17157/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T07:05:20Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"361mg7HCpdZbaDU5cYLfR74vRV8oTGe+LhwQ2rF0GfkilC2NX51s1G+//jVloRtRzEgdDdqzM7X05EGWBk0uCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-14T23:14:50.762924Z"},"content_sha256":"6c3f929f82b9864891c8fe1abda5b9a94216f21f87518446e7bec81ff60fdf6e","schema_version":"1.0","event_id":"sha256:6c3f929f82b9864891c8fe1abda5b9a94216f21f87518446e7bec81ff60fdf6e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/CCYRHP5GUMOWSPXXZN4ZQKQPTW/bundle.json","state_url":"https://pith.science/pith/CCYRHP5GUMOWSPXXZN4ZQKQPTW/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/CCYRHP5GUMOWSPXXZN4ZQKQPTW/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":"2026-08-14T23:14:50Z","links":{"resolver":"https://pith.science/pith/CCYRHP5GUMOWSPXXZN4ZQKQPTW","bundle":"https://pith.science/pith/CCYRHP5GUMOWSPXXZN4ZQKQPTW/bundle.json","state":"https://pith.science/pith/CCYRHP5GUMOWSPXXZN4ZQKQPTW/state.json","well_known_bundle":"https://pith.science/.well-known/pith/CCYRHP5GUMOWSPXXZN4ZQKQPTW/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:CCYRHP5GUMOWSPXXZN4ZQKQPTW","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":"8f2c97e7e1e8daac2c8daec1e2dacdfeff9af44b31b93bac054bca6b39575a05","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-10-26T05:01:09Z","title_canon_sha256":"a2373d8298a6249c1540fe27bca747e5ad6d86b0090ca87607785be58ddcdac1"},"schema_version":"1.0","source":{"id":"2310.17157","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2310.17157","created_at":"2026-07-05T07:05:20Z"},{"alias_kind":"arxiv_version","alias_value":"2310.17157v1","created_at":"2026-07-05T07:05:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2310.17157","created_at":"2026-07-05T07:05:20Z"},{"alias_kind":"pith_short_12","alias_value":"CCYRHP5GUMOW","created_at":"2026-07-05T07:05:20Z"},{"alias_kind":"pith_short_16","alias_value":"CCYRHP5GUMOWSPXX","created_at":"2026-07-05T07:05:20Z"},{"alias_kind":"pith_short_8","alias_value":"CCYRHP5G","created_at":"2026-07-05T07:05:20Z"}],"graph_snapshots":[{"event_id":"sha256:6c3f929f82b9864891c8fe1abda5b9a94216f21f87518446e7bec81ff60fdf6e","target":"graph","created_at":"2026-07-05T07:05:20Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2310.17157/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) with hundreds of billions of parameters have sparked a new wave of exciting AI applications. However, they are computationally expensive at inference time. Sparsity is a natural approach to reduce this cost, but existing methods either require costly retraining, have to forgo LLM's in-context learning ability, or do not yield wall-clock time speedup on modern hardware. We hypothesize that contextual sparsity, which are small, input-dependent sets of attention heads and MLP parameters that yield approximately the same output as the dense model for a given input, can","authors_text":"Anshumali Shrivastava, Beidi Chen, Binhang Yuan, Ce Zhang, Christopher Re, Jue Wang, Tianyi Zhou, Tri Dao, Yuandong Tian, Zhao Song, Zichang Liu","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-10-26T05:01:09Z","title":"Deja Vu: Contextual Sparsity for Efficient LLMs at Inference Time"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2310.17157","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:4e4c540998288a4fee14bed27f9531c087f4176c24b34cd96c03b532a4718a2a","target":"record","created_at":"2026-07-05T07:05:20Z","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":"8f2c97e7e1e8daac2c8daec1e2dacdfeff9af44b31b93bac054bca6b39575a05","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-10-26T05:01:09Z","title_canon_sha256":"a2373d8298a6249c1540fe27bca747e5ad6d86b0090ca87607785be58ddcdac1"},"schema_version":"1.0","source":{"id":"2310.17157","kind":"arxiv","version":1}},"canonical_sha256":"10b113bfa6a31d693ef7cb79982a0f9d86b27afe8ab7b2aac828535a68a5f554","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"10b113bfa6a31d693ef7cb79982a0f9d86b27afe8ab7b2aac828535a68a5f554","first_computed_at":"2026-07-05T07:05:20.760782Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T07:05:20.760782Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Tsm1ehF9q1E3BlHgfNFekmUmDOzfLJXS5S/fUf9Pr3n64O/prM3B4+rcz614L3lPNChliR78ptHUz886luHcCw==","signature_status":"signed_v1","signed_at":"2026-07-05T07:05:20.761187Z","signed_message":"canonical_sha256_bytes"},"source_id":"2310.17157","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4e4c540998288a4fee14bed27f9531c087f4176c24b34cd96c03b532a4718a2a","sha256:6c3f929f82b9864891c8fe1abda5b9a94216f21f87518446e7bec81ff60fdf6e"],"state_sha256":"7ba091d10f033c23783f96a7633c2f65e121a3da2e2dd9b86d932c419d18660f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dBII6mQ69EyhC8u/5LrUaB0lIcPfOvy9Lx+0NLuwc4yewyLiHGr2VvZUWEhN8UXfXfz5LpI0tKpr2thgINRiBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-14T23:14:50.770202Z","bundle_sha256":"bd6cdba542e5c669742d82313e25d20b94221a07f55ec1a57cce0a475252a500"}}