{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:TOY7SBLOTKY7LWUY7F2XUIZQGH","short_pith_number":"pith:TOY7SBLO","canonical_record":{"source":{"id":"2505.18086","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-23T16:43:03Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"e8ea41df1fae72636a80cb30e15121cdf41d60223dcd6a44c104c6564d2acb96","abstract_canon_sha256":"57eacd0c321acaaee2aee6e21fe521ec24eba437dc7f4177d8f5cc3794a3b064"},"schema_version":"1.0"},"canonical_sha256":"9bb1f9056e9ab1f5da98f9757a233031f4eae81c8876f0b94ffb846f26b444a1","source":{"kind":"arxiv","id":"2505.18086","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.18086","created_at":"2026-07-05T11:08:36Z"},{"alias_kind":"arxiv_version","alias_value":"2505.18086v1","created_at":"2026-07-05T11:08:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.18086","created_at":"2026-07-05T11:08:36Z"},{"alias_kind":"pith_short_12","alias_value":"TOY7SBLOTKY7","created_at":"2026-07-05T11:08:36Z"},{"alias_kind":"pith_short_16","alias_value":"TOY7SBLOTKY7LWUY","created_at":"2026-07-05T11:08:36Z"},{"alias_kind":"pith_short_8","alias_value":"TOY7SBLO","created_at":"2026-07-05T11:08:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:TOY7SBLOTKY7LWUY7F2XUIZQGH","target":"record","payload":{"canonical_record":{"source":{"id":"2505.18086","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-23T16:43:03Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"e8ea41df1fae72636a80cb30e15121cdf41d60223dcd6a44c104c6564d2acb96","abstract_canon_sha256":"57eacd0c321acaaee2aee6e21fe521ec24eba437dc7f4177d8f5cc3794a3b064"},"schema_version":"1.0"},"canonical_sha256":"9bb1f9056e9ab1f5da98f9757a233031f4eae81c8876f0b94ffb846f26b444a1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:08:36.465580Z","signature_b64":"p5zmwjo1H5q/DIv38eRebp8HJOSu8khf8yWf2hXvXWqEo0wcG0HVdXz76t3Ns7AF3BfjSO9C3f4Gt72ccErYDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9bb1f9056e9ab1f5da98f9757a233031f4eae81c8876f0b94ffb846f26b444a1","last_reissued_at":"2026-07-05T11:08:36.465033Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:08:36.465033Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2505.18086","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:08:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Js6fajuE7pMMZR8p1tN2hgnQcrWzCPyDeAj9/YrTE0DPj5JoaB/Y+g6pH4oVU5LbLuUcWdFuQ7byBRoPRr3OCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T22:44:17.793669Z"},"content_sha256":"4091ad4ef9c52e70f3209ca18d9de62c9e07b6c775aece390f5bfc44e50a70bc","schema_version":"1.0","event_id":"sha256:4091ad4ef9c52e70f3209ca18d9de62c9e07b6c775aece390f5bfc44e50a70bc"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:TOY7SBLOTKY7LWUY7F2XUIZQGH","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Stable Reinforcement Learning for Efficient Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.AI","authors_text":"Muzhi Dai, Qingyi Si, Shixuan Liu","submitted_at":"2025-05-23T16:43:03Z","abstract_excerpt":"The success of Deepseek-R1 has drawn the LLM community's attention to reinforcement learning (RL) methods like GRPO. However, such rule-based 0/1 outcome reward methods lack the capability to regulate the intermediate reasoning processes during chain-of-thought (CoT) generation, leading to severe overthinking phenomena. In response, recent studies have designed reward functions to reinforce models' behaviors in producing shorter yet correct completions. Nevertheless, we observe that these length-penalty reward functions exacerbate RL training instability: as the completion length decreases, mo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.18086","kind":"arxiv","version":1},"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/2505.18086/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:08:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QgEDsT/N+GyTb87N3DtL2YROiu/ATnjIT3hBJfiT/ABK/7Krba//OtrXWopxLFRZ7Sk7zglBIDHMZ3Dil/dgAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T22:44:17.794198Z"},"content_sha256":"6fa8c258b33a4ab0774302d92bbd81cc47675bb4da674c1da030c793107fd456","schema_version":"1.0","event_id":"sha256:6fa8c258b33a4ab0774302d92bbd81cc47675bb4da674c1da030c793107fd456"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/TOY7SBLOTKY7LWUY7F2XUIZQGH/bundle.json","state_url":"https://pith.science/pith/TOY7SBLOTKY7LWUY7F2XUIZQGH/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/TOY7SBLOTKY7LWUY7F2XUIZQGH/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-07T22:44:17Z","links":{"resolver":"https://pith.science/pith/TOY7SBLOTKY7LWUY7F2XUIZQGH","bundle":"https://pith.science/pith/TOY7SBLOTKY7LWUY7F2XUIZQGH/bundle.json","state":"https://pith.science/pith/TOY7SBLOTKY7LWUY7F2XUIZQGH/state.json","well_known_bundle":"https://pith.science/.well-known/pith/TOY7SBLOTKY7LWUY7F2XUIZQGH/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:TOY7SBLOTKY7LWUY7F2XUIZQGH","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"57eacd0c321acaaee2aee6e21fe521ec24eba437dc7f4177d8f5cc3794a3b064","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-23T16:43:03Z","title_canon_sha256":"e8ea41df1fae72636a80cb30e15121cdf41d60223dcd6a44c104c6564d2acb96"},"schema_version":"1.0","source":{"id":"2505.18086","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2505.18086","created_at":"2026-07-05T11:08:36Z"},{"alias_kind":"arxiv_version","alias_value":"2505.18086v1","created_at":"2026-07-05T11:08:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.18086","created_at":"2026-07-05T11:08:36Z"},{"alias_kind":"pith_short_12","alias_value":"TOY7SBLOTKY7","created_at":"2026-07-05T11:08:36Z"},{"alias_kind":"pith_short_16","alias_value":"TOY7SBLOTKY7LWUY","created_at":"2026-07-05T11:08:36Z"},{"alias_kind":"pith_short_8","alias_value":"TOY7SBLO","created_at":"2026-07-05T11:08:36Z"}],"graph_snapshots":[{"event_id":"sha256:6fa8c258b33a4ab0774302d92bbd81cc47675bb4da674c1da030c793107fd456","target":"graph","created_at":"2026-07-05T11:08:36Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2505.18086/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The success of Deepseek-R1 has drawn the LLM community's attention to reinforcement learning (RL) methods like GRPO. However, such rule-based 0/1 outcome reward methods lack the capability to regulate the intermediate reasoning processes during chain-of-thought (CoT) generation, leading to severe overthinking phenomena. In response, recent studies have designed reward functions to reinforce models' behaviors in producing shorter yet correct completions. Nevertheless, we observe that these length-penalty reward functions exacerbate RL training instability: as the completion length decreases, mo","authors_text":"Muzhi Dai, Qingyi Si, Shixuan Liu","cross_cats":["cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-23T16:43:03Z","title":"Stable Reinforcement Learning for Efficient Reasoning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.18086","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:4091ad4ef9c52e70f3209ca18d9de62c9e07b6c775aece390f5bfc44e50a70bc","target":"record","created_at":"2026-07-05T11:08:36Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"57eacd0c321acaaee2aee6e21fe521ec24eba437dc7f4177d8f5cc3794a3b064","cross_cats_sorted":["cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-05-23T16:43:03Z","title_canon_sha256":"e8ea41df1fae72636a80cb30e15121cdf41d60223dcd6a44c104c6564d2acb96"},"schema_version":"1.0","source":{"id":"2505.18086","kind":"arxiv","version":1}},"canonical_sha256":"9bb1f9056e9ab1f5da98f9757a233031f4eae81c8876f0b94ffb846f26b444a1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9bb1f9056e9ab1f5da98f9757a233031f4eae81c8876f0b94ffb846f26b444a1","first_computed_at":"2026-07-05T11:08:36.465033Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:08:36.465033Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"p5zmwjo1H5q/DIv38eRebp8HJOSu8khf8yWf2hXvXWqEo0wcG0HVdXz76t3Ns7AF3BfjSO9C3f4Gt72ccErYDQ==","signature_status":"signed_v1","signed_at":"2026-07-05T11:08:36.465580Z","signed_message":"canonical_sha256_bytes"},"source_id":"2505.18086","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4091ad4ef9c52e70f3209ca18d9de62c9e07b6c775aece390f5bfc44e50a70bc","sha256:6fa8c258b33a4ab0774302d92bbd81cc47675bb4da674c1da030c793107fd456"],"state_sha256":"aa467ba6752da12906bbb2b86a318cad2ba23cf5f243e507c254bf9f57ce8b6a"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"GQBFVAaXmfGsl3alieFrFWgJm8MsHAFRWOV2sG8qY2h3kXu1si2ziM7ZYa4GWS/X+w/RgN8PRE9Z244VRPT5Bw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T22:44:17.798021Z","bundle_sha256":"c7cdcb49f398b4781a42be4d3f5a45b7b82a9bf5ad93d7c72ca1d33f7c792698"}}