{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:JQX5JYALO6BJRB6LLMSF7UQXKH","short_pith_number":"pith:JQX5JYAL","schema_version":"1.0","canonical_sha256":"4c2fd4e00b77829887cb5b245fd21751d85163c18b6ae0ad711245800a6a03a6","source":{"kind":"arxiv","id":"2502.12134","version":2},"attestation_state":"computed","paper":{"title":"SoftCoT: Soft Chain-of-Thought for Efficient Reasoning with LLMs","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chunyan Miao, Xu Guo, Yige Xu, Zhiwei Zeng","submitted_at":"2025-02-17T18:52:29Z","abstract_excerpt":"Chain-of-Thought (CoT) reasoning enables Large Language Models (LLMs) to solve complex reasoning tasks by generating intermediate reasoning steps. However, most existing approaches focus on hard token decoding, which constrains reasoning within the discrete vocabulary space and may not always be optimal. While recent efforts explore continuous-space reasoning, they often require full-model fine-tuning and suffer from catastrophic forgetting, limiting their applicability to state-of-the-art LLMs that already perform well in zero-shot settings with a proper instruction. To address this challenge"},"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":"2502.12134","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-17T18:52:29Z","cross_cats_sorted":[],"title_canon_sha256":"b4e82c728ce1d56b2acb2e3270121a0c39a81361517b73d89e059d04c1f1a427","abstract_canon_sha256":"e3b6974b13fa0166f9dc44652ff8f8dcd55aa6c345a9d7079b2e88eb4fbda2b5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:10:35.235382Z","signature_b64":"C+EVpqmK7giDbyznZjL6e/NiqCWXRyUeF74v8U5mOf+SwXNEzwK733FcVmC8d55JPnD9GfyDBvUFal3a/1rJBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4c2fd4e00b77829887cb5b245fd21751d85163c18b6ae0ad711245800a6a03a6","last_reissued_at":"2026-07-05T11:10:35.234748Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:10:35.234748Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"SoftCoT: Soft Chain-of-Thought for Efficient Reasoning with LLMs","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chunyan Miao, Xu Guo, Yige Xu, Zhiwei Zeng","submitted_at":"2025-02-17T18:52:29Z","abstract_excerpt":"Chain-of-Thought (CoT) reasoning enables Large Language Models (LLMs) to solve complex reasoning tasks by generating intermediate reasoning steps. However, most existing approaches focus on hard token decoding, which constrains reasoning within the discrete vocabulary space and may not always be optimal. While recent efforts explore continuous-space reasoning, they often require full-model fine-tuning and suffer from catastrophic forgetting, limiting their applicability to state-of-the-art LLMs that already perform well in zero-shot settings with a proper instruction. To address this challenge"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.12134","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/2502.12134/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":"2502.12134","created_at":"2026-07-05T11:10:35.234818+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.12134v2","created_at":"2026-07-05T11:10:35.234818+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.12134","created_at":"2026-07-05T11:10:35.234818+00:00"},{"alias_kind":"pith_short_12","alias_value":"JQX5JYALO6BJ","created_at":"2026-07-05T11:10:35.234818+00:00"},{"alias_kind":"pith_short_16","alias_value":"JQX5JYALO6BJRB6L","created_at":"2026-07-05T11:10:35.234818+00:00"},{"alias_kind":"pith_short_8","alias_value":"JQX5JYAL","created_at":"2026-07-05T11:10:35.234818+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":14,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.02248","citing_title":"Geometric Latent Reasoning Induces Shorter Generations in LLMs","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01168","citing_title":"Thinking Economically: A Hierarchical Framework for Adaptive-Complexity Reasoning in LLMs","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09563","citing_title":"PRISM: Recovering Instruction Sets from Language Model Activations","ref_index":52,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21951","citing_title":"Dynamic Mixture of Latent Memories for Self-Evolving Agents","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2602.08324","citing_title":"Towards Efficient Large Language Reasoning Models via Extreme-Ratio Chain-of-Thought Compression","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2603.17837","citing_title":"The Silent Thought: Modeling Internal Cognition in Full-Duplex Spoken Dialogue Models via Latent Reasoning","ref_index":37,"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":35,"is_internal_anchor":false},{"citing_arxiv_id":"2601.19917","citing_title":"PILOT: Planning via Internalized Latent Optimization Trajectories for Large Language Models","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2601.06803","citing_title":"Forest Before Trees: Latent Superposition for Efficient Visual Reasoning","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2603.17837","citing_title":"The Silent Thought: Modeling Internal Cognition in Full-Duplex Spoken Dialogue Models via Latent Reasoning","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2503.16419","citing_title":"Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models","ref_index":206,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06165","citing_title":"Post Reasoning: Improving the Performance of Non-Thinking Models at No Cost","ref_index":90,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08299","citing_title":"SeLaR: Selective Latent Reasoning in Large Language Models","ref_index":50,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21027","citing_title":"HypEHR: Hyperbolic Modeling of Electronic Health Records for Efficient Question Answering","ref_index":23,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JQX5JYALO6BJRB6LLMSF7UQXKH","json":"https://pith.science/pith/JQX5JYALO6BJRB6LLMSF7UQXKH.json","graph_json":"https://pith.science/api/pith-number/JQX5JYALO6BJRB6LLMSF7UQXKH/graph.json","events_json":"https://pith.science/api/pith-number/JQX5JYALO6BJRB6LLMSF7UQXKH/events.json","paper":"https://pith.science/paper/JQX5JYAL"},"agent_actions":{"view_html":"https://pith.science/pith/JQX5JYALO6BJRB6LLMSF7UQXKH","download_json":"https://pith.science/pith/JQX5JYALO6BJRB6LLMSF7UQXKH.json","view_paper":"https://pith.science/paper/JQX5JYAL","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.12134&json=true","fetch_graph":"https://pith.science/api/pith-number/JQX5JYALO6BJRB6LLMSF7UQXKH/graph.json","fetch_events":"https://pith.science/api/pith-number/JQX5JYALO6BJRB6LLMSF7UQXKH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JQX5JYALO6BJRB6LLMSF7UQXKH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JQX5JYALO6BJRB6LLMSF7UQXKH/action/storage_attestation","attest_author":"https://pith.science/pith/JQX5JYALO6BJRB6LLMSF7UQXKH/action/author_attestation","sign_citation":"https://pith.science/pith/JQX5JYALO6BJRB6LLMSF7UQXKH/action/citation_signature","submit_replication":"https://pith.science/pith/JQX5JYALO6BJRB6LLMSF7UQXKH/action/replication_record"}},"created_at":"2026-07-05T11:10:35.234818+00:00","updated_at":"2026-07-05T11:10:35.234818+00:00"}