{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:755WHYYQTX2FUFK5PDYWDEY5N3","short_pith_number":"pith:755WHYYQ","schema_version":"1.0","canonical_sha256":"ff7b63e3109df45a155d78f161931d6efad4f136f16268088129b4de0dec0879","source":{"kind":"arxiv","id":"2507.06203","version":2},"attestation_state":"computed","paper":{"title":"A Survey on Latent Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andrew Smith, Assel Kembay, Binh Nguyen, Chenghua Lin, Chongxuan Li, Chujie Zheng, Dawei Zhu, Ge Zhang, Hao Wang, Jason Eshraghian, Jiaheng Liu, Jian Yang, Jinfa Huang, Kaiwen Xue, Rui-Jie Zhu, Taylor Kergan, Tianhao Cheng, Tianhao Peng, Tianle Cai, Tianyu Liu, Wangchunshu Zhou, Wenhao Huang, Xingwei Qu, Xuanliang Zhang, Yongchi Zhao, Yong Shan, Yuhong Chou, Yuqi Pan, Yuyin Zhou, Zefan Cai, Zhaoxiang Zhang, Zhenhe Wu, Zhoujun Li","submitted_at":"2025-07-08T17:29:07Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated impressive reasoning capabilities, especially when guided by explicit chain-of-thought (CoT) reasoning that verbalizes intermediate steps. While CoT improves both interpretability and accuracy, its dependence on natural language reasoning limits the model's expressive bandwidth. Latent reasoning tackles this bottleneck by performing multi-step inference entirely in the model's continuous hidden state, eliminating token-level supervision. To advance latent reasoning research, this survey provides a comprehensive overview of the emerging field of la"},"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":"2507.06203","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-08T17:29:07Z","cross_cats_sorted":[],"title_canon_sha256":"8c5511994f90c511b6eb97f396751452d97d3e3debdc5afc6880b0d6220a822f","abstract_canon_sha256":"ef6e54afdc8ca38552b0f2a400a0cc4001da1b8377aa3708e727f8fa39037b66"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:34:53.960758Z","signature_b64":"e56SDd1kdU83YjeSHTz8EcQeLSDWQ3Fgp+0QrX9hVk8eKKeQpDHlvkKktgU9rqxdkhuJaQj5vHfASa7iq3vcBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ff7b63e3109df45a155d78f161931d6efad4f136f16268088129b4de0dec0879","last_reissued_at":"2026-07-05T11:34:53.960276Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:34:53.960276Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Survey on Latent Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Andrew Smith, Assel Kembay, Binh Nguyen, Chenghua Lin, Chongxuan Li, Chujie Zheng, Dawei Zhu, Ge Zhang, Hao Wang, Jason Eshraghian, Jiaheng Liu, Jian Yang, Jinfa Huang, Kaiwen Xue, Rui-Jie Zhu, Taylor Kergan, Tianhao Cheng, Tianhao Peng, Tianle Cai, Tianyu Liu, Wangchunshu Zhou, Wenhao Huang, Xingwei Qu, Xuanliang Zhang, Yongchi Zhao, Yong Shan, Yuhong Chou, Yuqi Pan, Yuyin Zhou, Zefan Cai, Zhaoxiang Zhang, Zhenhe Wu, Zhoujun Li","submitted_at":"2025-07-08T17:29:07Z","abstract_excerpt":"Large Language Models (LLMs) have demonstrated impressive reasoning capabilities, especially when guided by explicit chain-of-thought (CoT) reasoning that verbalizes intermediate steps. While CoT improves both interpretability and accuracy, its dependence on natural language reasoning limits the model's expressive bandwidth. Latent reasoning tackles this bottleneck by performing multi-step inference entirely in the model's continuous hidden state, eliminating token-level supervision. To advance latent reasoning research, this survey provides a comprehensive overview of the emerging field of la"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.06203","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/2507.06203/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":"2507.06203","created_at":"2026-07-05T11:34:53.960339+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.06203v2","created_at":"2026-07-05T11:34:53.960339+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.06203","created_at":"2026-07-05T11:34:53.960339+00:00"},{"alias_kind":"pith_short_12","alias_value":"755WHYYQTX2F","created_at":"2026-07-05T11:34:53.960339+00:00"},{"alias_kind":"pith_short_16","alias_value":"755WHYYQTX2FUFK5","created_at":"2026-07-05T11:34:53.960339+00:00"},{"alias_kind":"pith_short_8","alias_value":"755WHYYQ","created_at":"2026-07-05T11:34:53.960339+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":19,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.21943","citing_title":"Modularized Reinforcement Learning on LLMs: From MDP Creation to Exploration and Learning","ref_index":284,"is_internal_anchor":false},{"citing_arxiv_id":"2606.13061","citing_title":"LaME: Learning to Think in Latent Space for Multimodal Embedding via Information Bottleneck","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07157","citing_title":"Think Fast: Estimating No-CoT Task-Completion Time Horizons of Frontier AI Models","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.27684","citing_title":"Intuition-Guided Latent Reasoning for LLM-Based Recommendation","ref_index":70,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26733","citing_title":"Stabilizing Recurrent Dynamics for Test-Time Scalable Latent Reasoning in Looped Language Models","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31779","citing_title":"Bridging the Gap Between Latent and Explicit Reasoning with Looped Transformers","ref_index":193,"is_internal_anchor":false},{"citing_arxiv_id":"2606.29928","citing_title":"Latent-CURE for Breast Cancer Diagnosis","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07157","citing_title":"Think Fast: Estimating No-CoT Task-Completion Time Horizons of Frontier AI Models","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20075","citing_title":"CopT: Contrastive On-Policy Thinking with Continuous Spaces for General and Agentic Reasoning","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2511.08983","citing_title":"SpiralThinker: Latent Reasoning through an Iterative Process with Text-Latent Interleaving","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2601.19917","citing_title":"PILOT: Planning via Internalized Latent Optimization Trajectories for Large Language Models","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2510.25741","citing_title":"Scaling Latent Reasoning via Looped Language Models","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11011","citing_title":"LoopUS: Recasting Pretrained LLMs into Looped Latent Refinement Models","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11301","citing_title":"LatentRouter: Can We Choose the Right Multimodal Model Before Seeing Its Answer?","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06285","citing_title":"LatentRAG: Latent Reasoning and Retrieval for Efficient Agentic RAG","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07315","citing_title":"LaTER: Efficient Test-Time Reasoning via Latent Exploration and Explicit Verification","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.15726","citing_title":"LLM Reasoning Is Latent, Not the Chain of Thought","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21027","citing_title":"HypEHR: Hyperbolic Modeling of Electronic Health Records for Efficient Question Answering","ref_index":110,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02442","citing_title":"Measuring AI Reasoning: A Guide for Researchers","ref_index":104,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/755WHYYQTX2FUFK5PDYWDEY5N3","json":"https://pith.science/pith/755WHYYQTX2FUFK5PDYWDEY5N3.json","graph_json":"https://pith.science/api/pith-number/755WHYYQTX2FUFK5PDYWDEY5N3/graph.json","events_json":"https://pith.science/api/pith-number/755WHYYQTX2FUFK5PDYWDEY5N3/events.json","paper":"https://pith.science/paper/755WHYYQ"},"agent_actions":{"view_html":"https://pith.science/pith/755WHYYQTX2FUFK5PDYWDEY5N3","download_json":"https://pith.science/pith/755WHYYQTX2FUFK5PDYWDEY5N3.json","view_paper":"https://pith.science/paper/755WHYYQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.06203&json=true","fetch_graph":"https://pith.science/api/pith-number/755WHYYQTX2FUFK5PDYWDEY5N3/graph.json","fetch_events":"https://pith.science/api/pith-number/755WHYYQTX2FUFK5PDYWDEY5N3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/755WHYYQTX2FUFK5PDYWDEY5N3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/755WHYYQTX2FUFK5PDYWDEY5N3/action/storage_attestation","attest_author":"https://pith.science/pith/755WHYYQTX2FUFK5PDYWDEY5N3/action/author_attestation","sign_citation":"https://pith.science/pith/755WHYYQTX2FUFK5PDYWDEY5N3/action/citation_signature","submit_replication":"https://pith.science/pith/755WHYYQTX2FUFK5PDYWDEY5N3/action/replication_record"}},"created_at":"2026-07-05T11:34:53.960339+00:00","updated_at":"2026-07-05T11:34:53.960339+00:00"}