{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BWCXFO5GNG6AXSFCK6YVSDNQG2","short_pith_number":"pith:BWCXFO5G","schema_version":"1.0","canonical_sha256":"0d8572bba669bc0bc8a257b1590db0369bef7b077bf494f6de4448203094df6d","source":{"kind":"arxiv","id":"2507.10628","version":2},"attestation_state":"computed","paper":{"title":"GHPO: Adaptive Guidance for Stable and Efficient LLM Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Cheng Gong, Dandan Tu, Qingfu Zhang, Ran Chen, Rui Liu, Shoubo Hu, Suiyun Zhang, Xinyu Fu, Yaofang Liu, Ziru Liu","submitted_at":"2025-07-14T08:10:00Z","abstract_excerpt":"Reinforcement Learning with Verifiable Rewards (RLVR) has recently emerged as a powerful paradigm for facilitating the self-improvement of large language models (LLMs), particularly in the domain of complex reasoning tasks. However, prevailing on-policy RL methods often contend with significant training instability and inefficiency. This is primarily due to a capacity-difficulty mismatch, where the complexity of training data frequently outpaces the model's current capabilities, leading to critically sparse reward signals and stalled learning progress. This challenge is particularly acute for "},"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":"2507.10628","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-07-14T08:10:00Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"fa2bdc12fbf36674e897298b2517d06c69c7195498b60e31be9ebc963c5dffa9","abstract_canon_sha256":"2a9a34f39ca0decba791a92f9733d4b9047bc2f6194ba1c13879dfc26979a3f2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:38:09.650540Z","signature_b64":"uZezRo0gYe3nXf+dzK5Ci88DoR90Y1fPho3hwKqzqTEumkEvQ/oGtPktLevHx/BhB+2tB3HnCy0dXL99NXBECg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0d8572bba669bc0bc8a257b1590db0369bef7b077bf494f6de4448203094df6d","last_reissued_at":"2026-07-05T11:38:09.649964Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:38:09.649964Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"GHPO: Adaptive Guidance for Stable and Efficient LLM Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Cheng Gong, Dandan Tu, Qingfu Zhang, Ran Chen, Rui Liu, Shoubo Hu, Suiyun Zhang, Xinyu Fu, Yaofang Liu, Ziru Liu","submitted_at":"2025-07-14T08:10:00Z","abstract_excerpt":"Reinforcement Learning with Verifiable Rewards (RLVR) has recently emerged as a powerful paradigm for facilitating the self-improvement of large language models (LLMs), particularly in the domain of complex reasoning tasks. However, prevailing on-policy RL methods often contend with significant training instability and inefficiency. This is primarily due to a capacity-difficulty mismatch, where the complexity of training data frequently outpaces the model's current capabilities, leading to critically sparse reward signals and stalled learning progress. This challenge is particularly acute for "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.10628","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/2507.10628/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":"2507.10628","created_at":"2026-07-05T11:38:09.650050+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.10628v2","created_at":"2026-07-05T11:38:09.650050+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.10628","created_at":"2026-07-05T11:38:09.650050+00:00"},{"alias_kind":"pith_short_12","alias_value":"BWCXFO5GNG6A","created_at":"2026-07-05T11:38:09.650050+00:00"},{"alias_kind":"pith_short_16","alias_value":"BWCXFO5GNG6AXSFC","created_at":"2026-07-05T11:38:09.650050+00:00"},{"alias_kind":"pith_short_8","alias_value":"BWCXFO5G","created_at":"2026-07-05T11:38:09.650050+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":12,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25450","citing_title":"The Generalization Spectrum: A Chromatographic Approach to Evaluating Learning Algorithms","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2606.25450","citing_title":"The Generalization Spectrum: A Chromatographic Approach to Evaluating Learning Algorithms","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03108","citing_title":"EvoTrainer: Co-Evolving LLM Policies and Training Harnesses for Autonomous Agentic Reinforcement Learning","ref_index":55,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25198","citing_title":"Hide to Guide: Learning via Semantic Masking","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22567","citing_title":"LANG: Reinforcement Learning for Multilingual Reasoning with Language-Adaptive Hint Guidance","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12969","citing_title":"Revisiting Reinforcement Learning with Verifiable Rewards from a Contrastive Perspective","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2509.02547","citing_title":"The Landscape of Agentic Reinforcement Learning for LLMs: A Survey","ref_index":59,"is_internal_anchor":false},{"citing_arxiv_id":"2601.06794","citing_title":"No More Stale Feedback: Co-Evolving Critics for Open-World Agent Learning","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2602.02350","citing_title":"Context Learning for Multi-Agent Discussion","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12969","citing_title":"Revisiting Reinforcement Learning with Verifiable Rewards from a Contrastive Perspective","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23318","citing_title":"Hidden States Know Where Reasoning Diverges: Credit Assignment via Span-Level Wasserstein Distance","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13017","citing_title":"PAL: Personal Adaptive Learner","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BWCXFO5GNG6AXSFCK6YVSDNQG2","json":"https://pith.science/pith/BWCXFO5GNG6AXSFCK6YVSDNQG2.json","graph_json":"https://pith.science/api/pith-number/BWCXFO5GNG6AXSFCK6YVSDNQG2/graph.json","events_json":"https://pith.science/api/pith-number/BWCXFO5GNG6AXSFCK6YVSDNQG2/events.json","paper":"https://pith.science/paper/BWCXFO5G"},"agent_actions":{"view_html":"https://pith.science/pith/BWCXFO5GNG6AXSFCK6YVSDNQG2","download_json":"https://pith.science/pith/BWCXFO5GNG6AXSFCK6YVSDNQG2.json","view_paper":"https://pith.science/paper/BWCXFO5G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.10628&json=true","fetch_graph":"https://pith.science/api/pith-number/BWCXFO5GNG6AXSFCK6YVSDNQG2/graph.json","fetch_events":"https://pith.science/api/pith-number/BWCXFO5GNG6AXSFCK6YVSDNQG2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BWCXFO5GNG6AXSFCK6YVSDNQG2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BWCXFO5GNG6AXSFCK6YVSDNQG2/action/storage_attestation","attest_author":"https://pith.science/pith/BWCXFO5GNG6AXSFCK6YVSDNQG2/action/author_attestation","sign_citation":"https://pith.science/pith/BWCXFO5GNG6AXSFCK6YVSDNQG2/action/citation_signature","submit_replication":"https://pith.science/pith/BWCXFO5GNG6AXSFCK6YVSDNQG2/action/replication_record"}},"created_at":"2026-07-05T11:38:09.650050+00:00","updated_at":"2026-07-05T11:38:09.650050+00:00"}