{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:KGPGLEZR2WAFQBPVL2KNJSHXYR","short_pith_number":"pith:KGPGLEZR","canonical_record":{"source":{"id":"2504.00927","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-04-01T15:59:32Z","cross_cats_sorted":[],"title_canon_sha256":"157fdd2541710c5b832ed3bdd628d001d41967775c2a96c39d6dfddbd6257a41","abstract_canon_sha256":"c107ba0531c9d551dc31c66bebc44ebcd6950e872ed18be4ad08bb3166e8a16a"},"schema_version":"1.0"},"canonical_sha256":"519e659331d5805805f55e94d4c8f7c47d540f4437eefd30134ee9e6a16c6e3c","source":{"kind":"arxiv","id":"2504.00927","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.00927","created_at":"2026-07-05T11:35:34Z"},{"alias_kind":"arxiv_version","alias_value":"2504.00927v2","created_at":"2026-07-05T11:35:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.00927","created_at":"2026-07-05T11:35:34Z"},{"alias_kind":"pith_short_12","alias_value":"KGPGLEZR2WAF","created_at":"2026-07-05T11:35:34Z"},{"alias_kind":"pith_short_16","alias_value":"KGPGLEZR2WAFQBPV","created_at":"2026-07-05T11:35:34Z"},{"alias_kind":"pith_short_8","alias_value":"KGPGLEZR","created_at":"2026-07-05T11:35:34Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:KGPGLEZR2WAFQBPVL2KNJSHXYR","target":"record","payload":{"canonical_record":{"source":{"id":"2504.00927","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-04-01T15:59:32Z","cross_cats_sorted":[],"title_canon_sha256":"157fdd2541710c5b832ed3bdd628d001d41967775c2a96c39d6dfddbd6257a41","abstract_canon_sha256":"c107ba0531c9d551dc31c66bebc44ebcd6950e872ed18be4ad08bb3166e8a16a"},"schema_version":"1.0"},"canonical_sha256":"519e659331d5805805f55e94d4c8f7c47d540f4437eefd30134ee9e6a16c6e3c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:35:34.328587Z","signature_b64":"2+5sZpIX4TsYAFNDgD7vSWs6yIRO9f6LjPw+q1dqPxRme5+q9IlwWrKDxaBG7oesDwgCowTIupBuAmHnNKgHCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"519e659331d5805805f55e94d4c8f7c47d540f4437eefd30134ee9e6a16c6e3c","last_reissued_at":"2026-07-05T11:35:34.328135Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:35:34.328135Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2504.00927","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:35:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xFN3xe68j5AaQPGqPYSS1Voz8qN8T+aL32ewlGMdff1vK+U/OKLoM7B3WY/X4x8OLxJC7FRNz/6kPMJyeKR3DA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T17:41:10.156178Z"},"content_sha256":"fea824d54e7a0add1400feac982b10acfe298f7b31ffd8286008609a9206f07e","schema_version":"1.0","event_id":"sha256:fea824d54e7a0add1400feac982b10acfe298f7b31ffd8286008609a9206f07e"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:KGPGLEZR2WAFQBPVL2KNJSHXYR","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Multi-Token Attention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jason Weston, Olga Golovneva, Sainbayar Sukhbaatar, Tianlu Wang","submitted_at":"2025-04-01T15:59:32Z","abstract_excerpt":"Soft attention is a critical mechanism powering LLMs to locate relevant parts within a given context. However, individual attention weights are determined by the similarity of only a single query and key token vector. This \"single token attention\" bottlenecks the amount of information used in distinguishing a relevant part from the rest of the context. To address this issue, we propose a new attention method, Multi-Token Attention (MTA), which allows LLMs to condition their attention weights on multiple query and key vectors simultaneously. This is achieved by applying convolution operations o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.00927","kind":"arxiv","version":2},"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/2504.00927/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":"2026-07-05T11:35:34Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Gv8j1cfQXL3HNpHrjK5Rn+WWWRNon+HkPcCLZIcZyL5fh/YMZBZBi6IBZhpVz9mAzjswnCiXP2o/5hY6tYHtBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T17:41:10.156677Z"},"content_sha256":"92d9d40e73b1e284755032e06f05fba7b995182f74618766197a74c19d8ccdaf","schema_version":"1.0","event_id":"sha256:92d9d40e73b1e284755032e06f05fba7b995182f74618766197a74c19d8ccdaf"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/KGPGLEZR2WAFQBPVL2KNJSHXYR/bundle.json","state_url":"https://pith.science/pith/KGPGLEZR2WAFQBPVL2KNJSHXYR/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/KGPGLEZR2WAFQBPVL2KNJSHXYR/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":"2026-08-08T17:41:10Z","links":{"resolver":"https://pith.science/pith/KGPGLEZR2WAFQBPVL2KNJSHXYR","bundle":"https://pith.science/pith/KGPGLEZR2WAFQBPVL2KNJSHXYR/bundle.json","state":"https://pith.science/pith/KGPGLEZR2WAFQBPVL2KNJSHXYR/state.json","well_known_bundle":"https://pith.science/.well-known/pith/KGPGLEZR2WAFQBPVL2KNJSHXYR/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:KGPGLEZR2WAFQBPVL2KNJSHXYR","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":"c107ba0531c9d551dc31c66bebc44ebcd6950e872ed18be4ad08bb3166e8a16a","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-04-01T15:59:32Z","title_canon_sha256":"157fdd2541710c5b832ed3bdd628d001d41967775c2a96c39d6dfddbd6257a41"},"schema_version":"1.0","source":{"id":"2504.00927","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.00927","created_at":"2026-07-05T11:35:34Z"},{"alias_kind":"arxiv_version","alias_value":"2504.00927v2","created_at":"2026-07-05T11:35:34Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.00927","created_at":"2026-07-05T11:35:34Z"},{"alias_kind":"pith_short_12","alias_value":"KGPGLEZR2WAF","created_at":"2026-07-05T11:35:34Z"},{"alias_kind":"pith_short_16","alias_value":"KGPGLEZR2WAFQBPV","created_at":"2026-07-05T11:35:34Z"},{"alias_kind":"pith_short_8","alias_value":"KGPGLEZR","created_at":"2026-07-05T11:35:34Z"}],"graph_snapshots":[{"event_id":"sha256:92d9d40e73b1e284755032e06f05fba7b995182f74618766197a74c19d8ccdaf","target":"graph","created_at":"2026-07-05T11:35:34Z","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/2504.00927/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Soft attention is a critical mechanism powering LLMs to locate relevant parts within a given context. However, individual attention weights are determined by the similarity of only a single query and key token vector. This \"single token attention\" bottlenecks the amount of information used in distinguishing a relevant part from the rest of the context. To address this issue, we propose a new attention method, Multi-Token Attention (MTA), which allows LLMs to condition their attention weights on multiple query and key vectors simultaneously. This is achieved by applying convolution operations o","authors_text":"Jason Weston, Olga Golovneva, Sainbayar Sukhbaatar, Tianlu Wang","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-04-01T15:59:32Z","title":"Multi-Token Attention"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.00927","kind":"arxiv","version":2},"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:fea824d54e7a0add1400feac982b10acfe298f7b31ffd8286008609a9206f07e","target":"record","created_at":"2026-07-05T11:35:34Z","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":"c107ba0531c9d551dc31c66bebc44ebcd6950e872ed18be4ad08bb3166e8a16a","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-04-01T15:59:32Z","title_canon_sha256":"157fdd2541710c5b832ed3bdd628d001d41967775c2a96c39d6dfddbd6257a41"},"schema_version":"1.0","source":{"id":"2504.00927","kind":"arxiv","version":2}},"canonical_sha256":"519e659331d5805805f55e94d4c8f7c47d540f4437eefd30134ee9e6a16c6e3c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"519e659331d5805805f55e94d4c8f7c47d540f4437eefd30134ee9e6a16c6e3c","first_computed_at":"2026-07-05T11:35:34.328135Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:35:34.328135Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"2+5sZpIX4TsYAFNDgD7vSWs6yIRO9f6LjPw+q1dqPxRme5+q9IlwWrKDxaBG7oesDwgCowTIupBuAmHnNKgHCw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:35:34.328587Z","signed_message":"canonical_sha256_bytes"},"source_id":"2504.00927","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:fea824d54e7a0add1400feac982b10acfe298f7b31ffd8286008609a9206f07e","sha256:92d9d40e73b1e284755032e06f05fba7b995182f74618766197a74c19d8ccdaf"],"state_sha256":"21ef589521b164e1b08702648ed649b352a7b4525127cd05ab06fbced5143960"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tC0XFsIlykHAwp07nFNwnb/ecUwfbCPgIbOdVnyJMdTQG+x5RvzTm8E+Q2PRxJYmorXn0VAvqIKV76WMRqyNBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T17:41:10.161393Z","bundle_sha256":"6d99ff6209216dea448a9465ff02bcd925663e876fdb9e95f51e676c59f9c318"}}