{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:LYKLUWUFF7EZ7WMJXSRYDDKKUM","short_pith_number":"pith:LYKLUWUF","schema_version":"1.0","canonical_sha256":"5e14ba5a852fc99fd989bca3818d4aa3155b58787f9b24d295801b8f09afdf61","source":{"kind":"arxiv","id":"2409.00088","version":2},"attestation_state":"computed","paper":{"title":"On-Device Language Models: A Comprehensive Review","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jiajun Xu, Qi Cai, Qun Wang, Wei Chen, Xin Gao, Zhiyuan Li, Ziyuan Ling","submitted_at":"2024-08-26T03:33:36Z","abstract_excerpt":"The advent of large language models (LLMs) revolutionized natural language processing applications, and running LLMs on edge devices has become increasingly attractive for reasons including reduced latency, data localization, and personalized user experiences. This comprehensive review examines the challenges of deploying computationally expensive LLMs on resource-constrained devices and explores innovative solutions across multiple domains. The paper investigates the development of on-device language models, their efficient architectures, including parameter sharing and modular designs, as we"},"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":"2409.00088","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-08-26T03:33:36Z","cross_cats_sorted":[],"title_canon_sha256":"13c66a3d6e5545ea806ae70724d8808acc91bacafd9bf7c31594f00afde64d81","abstract_canon_sha256":"f995c47f37028270d8a30398024b633a5c6b7babc4189c5526b20609f6c52154"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:07:02.750007Z","signature_b64":"XtrHFC4dqR7qJw97nHVkTHQIJMUoyA3cuHJePLLJXrY2/VTPPDppciDLryFXpVcv9LXNTTMulmmxBrW6TLB5CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5e14ba5a852fc99fd989bca3818d4aa3155b58787f9b24d295801b8f09afdf61","last_reissued_at":"2026-07-05T09:07:02.749461Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:07:02.749461Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On-Device Language Models: A Comprehensive Review","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Jiajun Xu, Qi Cai, Qun Wang, Wei Chen, Xin Gao, Zhiyuan Li, Ziyuan Ling","submitted_at":"2024-08-26T03:33:36Z","abstract_excerpt":"The advent of large language models (LLMs) revolutionized natural language processing applications, and running LLMs on edge devices has become increasingly attractive for reasons including reduced latency, data localization, and personalized user experiences. This comprehensive review examines the challenges of deploying computationally expensive LLMs on resource-constrained devices and explores innovative solutions across multiple domains. The paper investigates the development of on-device language models, their efficient architectures, including parameter sharing and modular designs, as we"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2409.00088","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/2409.00088/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":"2409.00088","created_at":"2026-07-05T09:07:02.749518+00:00"},{"alias_kind":"arxiv_version","alias_value":"2409.00088v2","created_at":"2026-07-05T09:07:02.749518+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2409.00088","created_at":"2026-07-05T09:07:02.749518+00:00"},{"alias_kind":"pith_short_12","alias_value":"LYKLUWUFF7EZ","created_at":"2026-07-05T09:07:02.749518+00:00"},{"alias_kind":"pith_short_16","alias_value":"LYKLUWUFF7EZ7WMJ","created_at":"2026-07-05T09:07:02.749518+00:00"},{"alias_kind":"pith_short_8","alias_value":"LYKLUWUF","created_at":"2026-07-05T09:07:02.749518+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":29,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.06371","citing_title":"The Impact of Security and Privacy Controls on Users' Emotional Engagement with Generative AI Chatbots","ref_index":86,"is_internal_anchor":true},{"citing_arxiv_id":"2606.25115","citing_title":"Forget to Improve: On-Device LLM-Agent Continual Learning via Budget-Curated Memory","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2606.23370","citing_title":"FlexServe: A Fast and Secure LLM Serving System for Mobile Devices with Flexible Resource Isolation","ref_index":74,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18147","citing_title":"WEQA: Wearable hEalth Question Answering with Query-Adaptive Agentic Reasoning","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2606.11257","citing_title":"Energy-Efficient On-Device RAG on a Mobile NPU: System Design and Benchmark on Snapdragon X Elite","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03248","citing_title":"Investigating Novice Researchers' Perceptions of Research Privacy Within LLM-Assisted Workflows","ref_index":110,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25698","citing_title":"Reference-Augmented Learning for Precise Tracking Policy of Tendon-Driven Continuum Robots","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03804","citing_title":"ScrapMem: A Bio-inspired Framework for On-device Personalized Agent Memory via Optical Forgetting","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2606.28843","citing_title":"The Heterogeneous Safety Impacts of Benign Multilingual Fine-Tuning","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29483","citing_title":"Fog Computing and Large Language Models: A vision for mutual beneficiaries","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2509.24730","citing_title":"Diamonds in the rough: Transforming SPARCs of imagination into a game concept by leveraging medium sized LLMs","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18831","citing_title":"Towards Discovery of Polymers for Insulin Delivery via Physics-Grounded Agentic Workflows","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18831","citing_title":"Towards Discovery of Polymers for Insulin Delivery via Physics-Grounded Agentic Workflows","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18535","citing_title":"Beyond Scaling: Agents Are Heading to the Edge","ref_index":67,"is_internal_anchor":false},{"citing_arxiv_id":"2510.02657","citing_title":"Less LLM, More Documents: Searching for Improved RAG","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2512.10371","citing_title":"AgentProg: Empowering Long-Horizon GUI Agents with Program-Guided Context Management","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2602.05902","citing_title":"CoreQ: Learning-Free Mismatch Correction and Successive Rounding for Quantization","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2603.09046","citing_title":"FlexServe: A Fast and Secure LLM Serving System for Mobile Devices with Flexible Resource Isolation","ref_index":68,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25699","citing_title":"NVLLM: A 3D NAND-Centric Architecture Enabling Edge on-Device LLM Inference","ref_index":33,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03804","citing_title":"ScrapMem: A Bio-inspired Framework for On-device Personalized Agent Memory via Optical Forgetting","ref_index":84,"is_internal_anchor":false},{"citing_arxiv_id":"2604.25975","citing_title":"Rethinking KV Cache Eviction via a Unified Information-Theoretic Objective","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22906","citing_title":"Network Edge Inference for Large Language Models: Principles, Techniques, and Opportunities","ref_index":176,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04726","citing_title":"RecGPT-Mobile: On-Device Large Language Models for User Intent Understanding in Taobao Feed Recommendation","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07725","citing_title":"SOD: Step-wise On-policy Distillation for Small Language Model Agents","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2604.17701","citing_title":"WISV: Wireless-Informed Semantic Verification for Distributed Speculative Decoding in Device-Edge LLM Inference","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LYKLUWUFF7EZ7WMJXSRYDDKKUM","json":"https://pith.science/pith/LYKLUWUFF7EZ7WMJXSRYDDKKUM.json","graph_json":"https://pith.science/api/pith-number/LYKLUWUFF7EZ7WMJXSRYDDKKUM/graph.json","events_json":"https://pith.science/api/pith-number/LYKLUWUFF7EZ7WMJXSRYDDKKUM/events.json","paper":"https://pith.science/paper/LYKLUWUF"},"agent_actions":{"view_html":"https://pith.science/pith/LYKLUWUFF7EZ7WMJXSRYDDKKUM","download_json":"https://pith.science/pith/LYKLUWUFF7EZ7WMJXSRYDDKKUM.json","view_paper":"https://pith.science/paper/LYKLUWUF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2409.00088&json=true","fetch_graph":"https://pith.science/api/pith-number/LYKLUWUFF7EZ7WMJXSRYDDKKUM/graph.json","fetch_events":"https://pith.science/api/pith-number/LYKLUWUFF7EZ7WMJXSRYDDKKUM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LYKLUWUFF7EZ7WMJXSRYDDKKUM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LYKLUWUFF7EZ7WMJXSRYDDKKUM/action/storage_attestation","attest_author":"https://pith.science/pith/LYKLUWUFF7EZ7WMJXSRYDDKKUM/action/author_attestation","sign_citation":"https://pith.science/pith/LYKLUWUFF7EZ7WMJXSRYDDKKUM/action/citation_signature","submit_replication":"https://pith.science/pith/LYKLUWUFF7EZ7WMJXSRYDDKKUM/action/replication_record"}},"created_at":"2026-07-05T09:07:02.749518+00:00","updated_at":"2026-07-05T09:07:02.749518+00:00"}