{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:UQUMPFH4EXMAQUL6UK35PVX3RV","short_pith_number":"pith:UQUMPFH4","schema_version":"1.0","canonical_sha256":"a428c794fc25d808517ea2b7d7d6fb8d5cfbb7747615c3089f4f73e5a1ccd88f","source":{"kind":"arxiv","id":"2502.21321","version":2},"attestation_state":"computed","paper":{"title":"LLM Post-Training: A Deep Dive into Reasoning Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CL","authors_text":"Fahad Shahbaz Khan, Hisham Cholakkal, Komal Kumar, Ming-Hsuan Yang, Mubarak Shah, Omkar Thawakar, Phillip H.S. Torr, Rao Muhammad Anwer, Salman Khan, Tajamul Ashraf","submitted_at":"2025-02-28T18:59:54Z","abstract_excerpt":"Large Language Models (LLMs) have transformed the natural language processing landscape and brought to life diverse applications. Pretraining on vast web-scale data has laid the foundation for these models, yet the research community is now increasingly shifting focus toward post-training techniques to achieve further breakthroughs. While pretraining provides a broad linguistic foundation, post-training methods enable LLMs to refine their knowledge, improve reasoning, enhance factual accuracy, and align more effectively with user intents and ethical considerations. Fine-tuning, reinforcement l"},"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":"2502.21321","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-02-28T18:59:54Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"a0e8e40c9ec3ea2aa5c083971ac7316fe29d2dafaca6c9d1cf25059764bf73f5","abstract_canon_sha256":"105a363c0c34b977d16effbe1de4f17447fef7efff57c6ba0a18da349662b12f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:37:45.409834Z","signature_b64":"dt2BjDxXU5N4CfnSNw0fWX6xYtzSEB1b5k5ofr+yyDG3uZK2KRz3hJGpcOJQ9Gewe+uT6Fnm7W5MSqUzxyK3Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a428c794fc25d808517ea2b7d7d6fb8d5cfbb7747615c3089f4f73e5a1ccd88f","last_reissued_at":"2026-07-05T10:37:45.409272Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:37:45.409272Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLM Post-Training: A Deep Dive into Reasoning Large Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CL","authors_text":"Fahad Shahbaz Khan, Hisham Cholakkal, Komal Kumar, Ming-Hsuan Yang, Mubarak Shah, Omkar Thawakar, Phillip H.S. Torr, Rao Muhammad Anwer, Salman Khan, Tajamul Ashraf","submitted_at":"2025-02-28T18:59:54Z","abstract_excerpt":"Large Language Models (LLMs) have transformed the natural language processing landscape and brought to life diverse applications. Pretraining on vast web-scale data has laid the foundation for these models, yet the research community is now increasingly shifting focus toward post-training techniques to achieve further breakthroughs. While pretraining provides a broad linguistic foundation, post-training methods enable LLMs to refine their knowledge, improve reasoning, enhance factual accuracy, and align more effectively with user intents and ethical considerations. Fine-tuning, reinforcement l"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.21321","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/2502.21321/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":"2502.21321","created_at":"2026-07-05T10:37:45.409346+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.21321v2","created_at":"2026-07-05T10:37:45.409346+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.21321","created_at":"2026-07-05T10:37:45.409346+00:00"},{"alias_kind":"pith_short_12","alias_value":"UQUMPFH4EXMA","created_at":"2026-07-05T10:37:45.409346+00:00"},{"alias_kind":"pith_short_16","alias_value":"UQUMPFH4EXMAQUL6","created_at":"2026-07-05T10:37:45.409346+00:00"},{"alias_kind":"pith_short_8","alias_value":"UQUMPFH4","created_at":"2026-07-05T10:37:45.409346+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":34,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.27369","citing_title":"Reinforcement Learning without Ground-Truth Solutions can Improve LLMs","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21943","citing_title":"Modularized Reinforcement Learning on LLMs: From MDP Creation to Exploration and Learning","ref_index":94,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18967","citing_title":"EfficientRollout: System-Aware Self-Speculative Decoding for RL Rollouts","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2606.13176","citing_title":"Mental-R1: Aligning LLM Reasoning for Mental Health Assessment","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12117","citing_title":"Soft-Prompt Tuning for Fair and Efficient LLM Benchmark Evaluation","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09396","citing_title":"PriFT: Prior-Support Guided Supervised Fine-Tuning","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09471","citing_title":"Escaping the KL Agreement Trap in On-Policy Distillation","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09092","citing_title":"From Shortcuts to Reasoning: Robust Post-Training of Theory of Mind with Reinforcement Learning","ref_index":66,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03077","citing_title":"Libra: Efficient Resource Management for Agentic RL Post-Training","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00047","citing_title":"Comprehensive AI governance requires addressing non-model gains","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20201","citing_title":"Long-Context Reasoning Through Proxy-Based Chain-of-Thought Tuning","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2504.02181","citing_title":"A Survey of Scaling in Large Language Model Reasoning","ref_index":88,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18721","citing_title":"General Preference Reinforcement Learning","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2601.21484","citing_title":"ETS: Energy-Guided Test-Time Scaling for Training-Free RL Alignment","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20201","citing_title":"Long-Context Reasoning Through Proxy-Based Chain-of-Thought Tuning","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18721","citing_title":"General Preference Reinforcement Learning","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18721","citing_title":"General Preference Reinforcement Learning","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2508.05015","citing_title":"SPaCe: Unlocking Sample-Efficient Large Language Models Training With Self-Pace Curriculum Learning","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2508.06412","citing_title":"Sample-efficient LLM Optimization with Reset Replay","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2601.21484","citing_title":"ETS: Energy-Guided Test-Time Scaling for Training-Free RL Alignment","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2504.13958","citing_title":"ToolRL: Reward is All Tool Learning Needs","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2604.03231","citing_title":"CoME-VL: Scaling Complementary Multi-Encoder Vision-Language Learning","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03295","citing_title":"Cheap Expertise: Mapping and Challenging Industry Perspectives in the Expert Data Gig Economy","ref_index":82,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08378","citing_title":"Reinforcement Learning for Scalable and Trustworthy Intelligent Systems","ref_index":199,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09608","citing_title":"Geometry Conflict: Explaining and Controlling Forgetting in LLM Continual Post-Training","ref_index":2,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UQUMPFH4EXMAQUL6UK35PVX3RV","json":"https://pith.science/pith/UQUMPFH4EXMAQUL6UK35PVX3RV.json","graph_json":"https://pith.science/api/pith-number/UQUMPFH4EXMAQUL6UK35PVX3RV/graph.json","events_json":"https://pith.science/api/pith-number/UQUMPFH4EXMAQUL6UK35PVX3RV/events.json","paper":"https://pith.science/paper/UQUMPFH4"},"agent_actions":{"view_html":"https://pith.science/pith/UQUMPFH4EXMAQUL6UK35PVX3RV","download_json":"https://pith.science/pith/UQUMPFH4EXMAQUL6UK35PVX3RV.json","view_paper":"https://pith.science/paper/UQUMPFH4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.21321&json=true","fetch_graph":"https://pith.science/api/pith-number/UQUMPFH4EXMAQUL6UK35PVX3RV/graph.json","fetch_events":"https://pith.science/api/pith-number/UQUMPFH4EXMAQUL6UK35PVX3RV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UQUMPFH4EXMAQUL6UK35PVX3RV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UQUMPFH4EXMAQUL6UK35PVX3RV/action/storage_attestation","attest_author":"https://pith.science/pith/UQUMPFH4EXMAQUL6UK35PVX3RV/action/author_attestation","sign_citation":"https://pith.science/pith/UQUMPFH4EXMAQUL6UK35PVX3RV/action/citation_signature","submit_replication":"https://pith.science/pith/UQUMPFH4EXMAQUL6UK35PVX3RV/action/replication_record"}},"created_at":"2026-07-05T10:37:45.409346+00:00","updated_at":"2026-07-05T10:37:45.409346+00:00"}