{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:IGVG3EXBRV6REICKQITQFQH3IF","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":"df57fb473746a54bc74dbde3e81d0127465394c86d0af2c8f27a5833187dd56a","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-16T23:15:09Z","title_canon_sha256":"801660522fa7d163aa430f7f68592a7e4d8e77810402b0a62f2e32746dbf1de6"},"schema_version":"1.0","source":{"id":"2504.12526","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.12526","created_at":"2026-07-05T10:50:33Z"},{"alias_kind":"arxiv_version","alias_value":"2504.12526v1","created_at":"2026-07-05T10:50:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.12526","created_at":"2026-07-05T10:50:33Z"},{"alias_kind":"pith_short_12","alias_value":"IGVG3EXBRV6R","created_at":"2026-07-05T10:50:33Z"},{"alias_kind":"pith_short_16","alias_value":"IGVG3EXBRV6REICK","created_at":"2026-07-05T10:50:33Z"},{"alias_kind":"pith_short_8","alias_value":"IGVG3EXB","created_at":"2026-07-05T10:50:33Z"}],"graph_snapshots":[{"event_id":"sha256:b8517f44a5539a8201c1675533bcac64bd17ffbef7751e1176facbbd7bda5ece","target":"graph","created_at":"2026-07-05T10:50:33Z","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/2504.12526/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Long-context language models exhibit impressive performance but remain challenging to deploy due to high GPU memory demands during inference. We propose Memory-efficient Offloaded Mini-sequence Inference (MOM), a method that partitions critical layers into smaller \"mini-sequences\" and integrates seamlessly with KV cache offloading. Experiments on various Llama, Qwen, and Mistral models demonstrate that MOM reduces peak memory usage by over 50\\% on average. On Meta-Llama-3.2-8B, MOM extends the maximum context length from 155k to 455k tokens on a single A100 80GB GPU, while keeping outputs iden","authors_text":"Anima Anandkumar, Cheng Luo, Junyang Zhang, Tianyi Zhu","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-16T23:15:09Z","title":"MOM: Memory-Efficient Offloaded Mini-Sequence Inference for Long Context Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.12526","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":"sha256:89cdfedad56d6555d9bf08ffb8bf9baa1ac579cf683f9245c5c2116a5f865fb6","target":"record","created_at":"2026-07-05T10:50:33Z","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":"df57fb473746a54bc74dbde3e81d0127465394c86d0af2c8f27a5833187dd56a","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-16T23:15:09Z","title_canon_sha256":"801660522fa7d163aa430f7f68592a7e4d8e77810402b0a62f2e32746dbf1de6"},"schema_version":"1.0","source":{"id":"2504.12526","kind":"arxiv","version":1}},"canonical_sha256":"41aa6d92e18d7d12204a822702c0fb41490d6330ac4a0982557e7828f3d4c313","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"41aa6d92e18d7d12204a822702c0fb41490d6330ac4a0982557e7828f3d4c313","first_computed_at":"2026-07-05T10:50:33.848433Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:50:33.848433Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"KD6h3l2sf00FYx9lVVysihWB4Eav+mKoi9S46afhS98b1+gpJhesZUt7mWMmsmv5mJMMAWNCidp9wwikb3QaAA==","signature_status":"signed_v1","signed_at":"2026-07-05T10:50:33.848923Z","signed_message":"canonical_sha256_bytes"},"source_id":"2504.12526","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:89cdfedad56d6555d9bf08ffb8bf9baa1ac579cf683f9245c5c2116a5f865fb6","sha256:b8517f44a5539a8201c1675533bcac64bd17ffbef7751e1176facbbd7bda5ece"],"state_sha256":"401f4ee293ebc733345ce7d0a5499bca2e536cc8e3d0810bdcb283eea17062f7"}