{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:UL5VIBMOKNTW5LP4EVIKLTMUT3","short_pith_number":"pith:UL5VIBMO","schema_version":"1.0","canonical_sha256":"a2fb54058e53676eadfc2550a5cd949ec7ee85e360a661f85e42f56e68f9f3c8","source":{"kind":"arxiv","id":"2411.04282","version":2},"attestation_state":"computed","paper":{"title":"Language Models are Hidden Reasoners: Unlocking Latent Reasoning Capabilities via Self-Rewarding","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.CL","cs.LG","stat.ML"],"primary_cat":"cs.AI","authors_text":"Akshara Prabhakar, Caiming Xiong, Haolin Chen, Huan Wang, Phil Mui, Ricky Ho, Shelby Heinecke, Silvio Savarese, Weiran Yao, Yihao Feng, Zuxin Liu","submitted_at":"2024-11-06T22:02:30Z","abstract_excerpt":"Large language models (LLMs) have shown impressive capabilities, but still struggle with complex reasoning tasks requiring multiple steps. While prompt-based methods like Chain-of-Thought (CoT) can improve LLM reasoning at inference time, optimizing reasoning capabilities during training remains challenging. We introduce LaTent Reasoning Optimization (LaTRO), a principled framework that formulates reasoning as sampling from a latent distribution and optimizes it via variational approaches. LaTRO enables LLMs to concurrently improve both their reasoning process and ability to evaluate reasoning"},"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":"2411.04282","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.AI","submitted_at":"2024-11-06T22:02:30Z","cross_cats_sorted":["cs.CL","cs.LG","stat.ML"],"title_canon_sha256":"3869e440608c97063823e157d21cd0a46caaa695876bc8ca17d5674421179a0d","abstract_canon_sha256":"2d09be8fd18d98489102277dd0721717223a068cad712ba468b692c58f06ff6d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:38:50.598330Z","signature_b64":"G+f9VfX/PxRcMbL1hhN3s9Y1fmSYP4wYnPPSkn8HhYwR6YFEeKjzFb+uguFcdB+O9isX0R16bmMBkoJQODFXAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a2fb54058e53676eadfc2550a5cd949ec7ee85e360a661f85e42f56e68f9f3c8","last_reissued_at":"2026-07-05T09:38:50.597816Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:38:50.597816Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Language Models are Hidden Reasoners: Unlocking Latent Reasoning Capabilities via Self-Rewarding","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.CL","cs.LG","stat.ML"],"primary_cat":"cs.AI","authors_text":"Akshara Prabhakar, Caiming Xiong, Haolin Chen, Huan Wang, Phil Mui, Ricky Ho, Shelby Heinecke, Silvio Savarese, Weiran Yao, Yihao Feng, Zuxin Liu","submitted_at":"2024-11-06T22:02:30Z","abstract_excerpt":"Large language models (LLMs) have shown impressive capabilities, but still struggle with complex reasoning tasks requiring multiple steps. While prompt-based methods like Chain-of-Thought (CoT) can improve LLM reasoning at inference time, optimizing reasoning capabilities during training remains challenging. We introduce LaTent Reasoning Optimization (LaTRO), a principled framework that formulates reasoning as sampling from a latent distribution and optimizes it via variational approaches. LaTRO enables LLMs to concurrently improve both their reasoning process and ability to evaluate reasoning"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.04282","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/2411.04282/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":"2411.04282","created_at":"2026-07-05T09:38:50.597877+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.04282v2","created_at":"2026-07-05T09:38:50.597877+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.04282","created_at":"2026-07-05T09:38:50.597877+00:00"},{"alias_kind":"pith_short_12","alias_value":"UL5VIBMOKNTW","created_at":"2026-07-05T09:38:50.597877+00:00"},{"alias_kind":"pith_short_16","alias_value":"UL5VIBMOKNTW5LP4","created_at":"2026-07-05T09:38:50.597877+00:00"},{"alias_kind":"pith_short_8","alias_value":"UL5VIBMO","created_at":"2026-07-05T09:38:50.597877+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.07663","citing_title":"Recursive Self-Improvement in AI: From Bounded Self-Refinement to Autonomous Research Loops","ref_index":101,"is_internal_anchor":true},{"citing_arxiv_id":"2606.27359","citing_title":"When are likely answers right? On Sequence Probability and Correctness in LLMs","ref_index":3,"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":20,"is_internal_anchor":false},{"citing_arxiv_id":"2606.19594","citing_title":"Unsupervised Causal Abstractions Discovery","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2602.12579","citing_title":"VI-CuRL: Stabilizing Verifier-Independent RL Reasoning via Confidence-Guided Variance Reduction","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2602.08167","citing_title":"Self-Supervised Bootstrapping of Action-Predictive Embodied Reasoning","ref_index":68,"is_internal_anchor":false},{"citing_arxiv_id":"2506.13351","citing_title":"Direct Reasoning Optimization: Token-Level Reasoning Reflectivity Meets Rubric Gates for Unverifiable Tasks","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2503.09567","citing_title":"Towards Reasoning Era: A Survey of Long Chain-of-Thought for Reasoning Large Language Models","ref_index":80,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10207","citing_title":"LASAR: Latent Adaptive Semantic Aligned Reasoning for Generative Recommendation","ref_index":3,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UL5VIBMOKNTW5LP4EVIKLTMUT3","json":"https://pith.science/pith/UL5VIBMOKNTW5LP4EVIKLTMUT3.json","graph_json":"https://pith.science/api/pith-number/UL5VIBMOKNTW5LP4EVIKLTMUT3/graph.json","events_json":"https://pith.science/api/pith-number/UL5VIBMOKNTW5LP4EVIKLTMUT3/events.json","paper":"https://pith.science/paper/UL5VIBMO"},"agent_actions":{"view_html":"https://pith.science/pith/UL5VIBMOKNTW5LP4EVIKLTMUT3","download_json":"https://pith.science/pith/UL5VIBMOKNTW5LP4EVIKLTMUT3.json","view_paper":"https://pith.science/paper/UL5VIBMO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.04282&json=true","fetch_graph":"https://pith.science/api/pith-number/UL5VIBMOKNTW5LP4EVIKLTMUT3/graph.json","fetch_events":"https://pith.science/api/pith-number/UL5VIBMOKNTW5LP4EVIKLTMUT3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UL5VIBMOKNTW5LP4EVIKLTMUT3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UL5VIBMOKNTW5LP4EVIKLTMUT3/action/storage_attestation","attest_author":"https://pith.science/pith/UL5VIBMOKNTW5LP4EVIKLTMUT3/action/author_attestation","sign_citation":"https://pith.science/pith/UL5VIBMOKNTW5LP4EVIKLTMUT3/action/citation_signature","submit_replication":"https://pith.science/pith/UL5VIBMOKNTW5LP4EVIKLTMUT3/action/replication_record"}},"created_at":"2026-07-05T09:38:50.597877+00:00","updated_at":"2026-07-05T09:38:50.597877+00:00"}