{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:LJIIUQDL7AKJXYKYQ3PZC4LSCG","short_pith_number":"pith:LJIIUQDL","canonical_record":{"source":{"id":"2607.16184","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-17T17:58:29Z","cross_cats_sorted":[],"title_canon_sha256":"8a0290799bd3c9328bd3ac65c779ceeee7b53d03fa44c34966a8732e3c93ff94","abstract_canon_sha256":"6dadc888a8af6039a91ed5e1ed6fd2bdc5b4421b4444ebbc538d033f70e10be4"},"schema_version":"1.0"},"canonical_sha256":"5a508a406bf8149be15886df91717211aad5e6061585fa8af4ed758211ad9647","source":{"kind":"arxiv","id":"2607.16184","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.16184","created_at":"2026-07-20T02:19:38Z"},{"alias_kind":"arxiv_version","alias_value":"2607.16184v1","created_at":"2026-07-20T02:19:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.16184","created_at":"2026-07-20T02:19:38Z"},{"alias_kind":"pith_short_12","alias_value":"LJIIUQDL7AKJ","created_at":"2026-07-20T02:19:38Z"},{"alias_kind":"pith_short_16","alias_value":"LJIIUQDL7AKJXYKY","created_at":"2026-07-20T02:19:38Z"},{"alias_kind":"pith_short_8","alias_value":"LJIIUQDL","created_at":"2026-07-20T02:19:38Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:LJIIUQDL7AKJXYKYQ3PZC4LSCG","target":"record","payload":{"canonical_record":{"source":{"id":"2607.16184","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-17T17:58:29Z","cross_cats_sorted":[],"title_canon_sha256":"8a0290799bd3c9328bd3ac65c779ceeee7b53d03fa44c34966a8732e3c93ff94","abstract_canon_sha256":"6dadc888a8af6039a91ed5e1ed6fd2bdc5b4421b4444ebbc538d033f70e10be4"},"schema_version":"1.0"},"canonical_sha256":"5a508a406bf8149be15886df91717211aad5e6061585fa8af4ed758211ad9647","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-20T02:19:38.362216Z","signature_b64":"/eSjcHNSAm6nw3TEKXKfdgdZYDPk/Xn+S0SXSnHEtIHzroPM+IueW462XRIUf2tHx1JrlhLU4gwkPE487V/BBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5a508a406bf8149be15886df91717211aad5e6061585fa8af4ed758211ad9647","last_reissued_at":"2026-07-20T02:19:38.361373Z","signature_status":"signed_v1","first_computed_at":"2026-07-20T02:19:38.361373Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2607.16184","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":"2026-07-20T02:19:38Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eZDHGllx/uYt4GJwOi11W5GMPul/RVN0MYORlCfiWeaGUkJfueUGT/vIjoAIEwHfzLKa5oZl3HURGEaU6qlLAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T09:08:49.922251Z"},"content_sha256":"f3a73e5eacdce61e69a5b7816100791398a162cc693b924386fa66f0ee816d41","schema_version":"1.0","event_id":"sha256:f3a73e5eacdce61e69a5b7816100791398a162cc693b924386fa66f0ee816d41"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:LJIIUQDL7AKJXYKYQ3PZC4LSCG","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"PagedWeight: Efficient MoE LLM Serving with Dynamic Quality-Aware Weight Quantization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Anisha Dasgupta, Sasa Misailovic, Yifan Zhao, Yuchen Yang","submitted_at":"2026-07-17T17:58:29Z","abstract_excerpt":"Mixture-of-Experts (MoE) is a popular class of large language models (LLMs), offering high efficiency and accuracy. However, in KV-cache-intensive serving scenarios, MoEs often exhibit a tension between the GPU memory requirements of the model weights and the growing KV cache. We propose PagedWeight, a novel management method for MoE LLM serving that dynamically quantizes MoE model's weights at runtime and balances expert-weight precision with the KV cache sizes. PagedWeight exposes and effectively navigates the complex tradeoff between the model's task accuracy, memory consumption, and throug"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.16184","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/2607.16184/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":"2026-07-20T02:19:38Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"AfErYxNOiaMWjbd1/4soh+J5kR7Yv93EaIkdXsIyZ2YLrYOvGAvvZyLXSf3iCqfcBTz+kWR9aP54O+nCvaOBCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T09:08:49.923199Z"},"content_sha256":"7b79b2e8fc71bf2f502472651b8df1fd0849aba4973b8a110c4304e21462ebc1","schema_version":"1.0","event_id":"sha256:7b79b2e8fc71bf2f502472651b8df1fd0849aba4973b8a110c4304e21462ebc1"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LJIIUQDL7AKJXYKYQ3PZC4LSCG/bundle.json","state_url":"https://pith.science/pith/LJIIUQDL7AKJXYKYQ3PZC4LSCG/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LJIIUQDL7AKJXYKYQ3PZC4LSCG/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":"2026-08-03T09:08:49Z","links":{"resolver":"https://pith.science/pith/LJIIUQDL7AKJXYKYQ3PZC4LSCG","bundle":"https://pith.science/pith/LJIIUQDL7AKJXYKYQ3PZC4LSCG/bundle.json","state":"https://pith.science/pith/LJIIUQDL7AKJXYKYQ3PZC4LSCG/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LJIIUQDL7AKJXYKYQ3PZC4LSCG/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:LJIIUQDL7AKJXYKYQ3PZC4LSCG","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":"6dadc888a8af6039a91ed5e1ed6fd2bdc5b4421b4444ebbc538d033f70e10be4","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-17T17:58:29Z","title_canon_sha256":"8a0290799bd3c9328bd3ac65c779ceeee7b53d03fa44c34966a8732e3c93ff94"},"schema_version":"1.0","source":{"id":"2607.16184","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2607.16184","created_at":"2026-07-20T02:19:38Z"},{"alias_kind":"arxiv_version","alias_value":"2607.16184v1","created_at":"2026-07-20T02:19:38Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.16184","created_at":"2026-07-20T02:19:38Z"},{"alias_kind":"pith_short_12","alias_value":"LJIIUQDL7AKJ","created_at":"2026-07-20T02:19:38Z"},{"alias_kind":"pith_short_16","alias_value":"LJIIUQDL7AKJXYKY","created_at":"2026-07-20T02:19:38Z"},{"alias_kind":"pith_short_8","alias_value":"LJIIUQDL","created_at":"2026-07-20T02:19:38Z"}],"graph_snapshots":[{"event_id":"sha256:7b79b2e8fc71bf2f502472651b8df1fd0849aba4973b8a110c4304e21462ebc1","target":"graph","created_at":"2026-07-20T02:19:38Z","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/2607.16184/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Mixture-of-Experts (MoE) is a popular class of large language models (LLMs), offering high efficiency and accuracy. However, in KV-cache-intensive serving scenarios, MoEs often exhibit a tension between the GPU memory requirements of the model weights and the growing KV cache. We propose PagedWeight, a novel management method for MoE LLM serving that dynamically quantizes MoE model's weights at runtime and balances expert-weight precision with the KV cache sizes. PagedWeight exposes and effectively navigates the complex tradeoff between the model's task accuracy, memory consumption, and throug","authors_text":"Anisha Dasgupta, Sasa Misailovic, Yifan Zhao, Yuchen Yang","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-17T17:58:29Z","title":"PagedWeight: Efficient MoE LLM Serving with Dynamic Quality-Aware Weight Quantization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.16184","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:f3a73e5eacdce61e69a5b7816100791398a162cc693b924386fa66f0ee816d41","target":"record","created_at":"2026-07-20T02:19:38Z","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":"6dadc888a8af6039a91ed5e1ed6fd2bdc5b4421b4444ebbc538d033f70e10be4","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-07-17T17:58:29Z","title_canon_sha256":"8a0290799bd3c9328bd3ac65c779ceeee7b53d03fa44c34966a8732e3c93ff94"},"schema_version":"1.0","source":{"id":"2607.16184","kind":"arxiv","version":1}},"canonical_sha256":"5a508a406bf8149be15886df91717211aad5e6061585fa8af4ed758211ad9647","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5a508a406bf8149be15886df91717211aad5e6061585fa8af4ed758211ad9647","first_computed_at":"2026-07-20T02:19:38.361373Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-20T02:19:38.361373Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"/eSjcHNSAm6nw3TEKXKfdgdZYDPk/Xn+S0SXSnHEtIHzroPM+IueW462XRIUf2tHx1JrlhLU4gwkPE487V/BBQ==","signature_status":"signed_v1","signed_at":"2026-07-20T02:19:38.362216Z","signed_message":"canonical_sha256_bytes"},"source_id":"2607.16184","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f3a73e5eacdce61e69a5b7816100791398a162cc693b924386fa66f0ee816d41","sha256:7b79b2e8fc71bf2f502472651b8df1fd0849aba4973b8a110c4304e21462ebc1"],"state_sha256":"0d85a1e08ca8b634c35200fdd02c70933381a13cd60ce5c9f9094b90355488ee"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SS5CFz7W5+6npMsRhNhmz6t+WUoHQsddbJmvo1QM70ls3YeZxKqT94yiZaqIlTrSb7//3l/lhJYQhrLigS+2Cw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-03T09:08:49.989811Z","bundle_sha256":"74c0c4abffc698a07088cea7f4e494c814361beec6de7f1488d0e825d12edaee"}}