{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:KIDQJHHECD3OXBRRZWSJ5NZJAU","short_pith_number":"pith:KIDQJHHE","schema_version":"1.0","canonical_sha256":"5207049ce410f6eb8631cda49eb729050489ccbe2e73b5786089cfefd7e6c764","source":{"kind":"arxiv","id":"2606.25757","version":1},"attestation_state":"computed","paper":{"title":"OPERA: Aligning Open-Ended Reasoning via Objective Perplexity-based Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Haoyang Dai, Hongming Tan, Ninghao Liu, Wenxuan Jiang, Xiaoyu Li, Xuecheng Wu, Xuezhi Cao, Zijian Zhang, Zining Fan","submitted_at":"2026-06-24T12:31:14Z","abstract_excerpt":"Reinforcement Learning (RL) has enabled LLMs to excel in objective reasoning tasks such as mathematics and code generation. However, applying RL to open-ended tasks, such as creative writing, remains challenging because LLM-as-a-judge reward models often exhibit stylistic biases and positional inconsistencies, leading to unstable supervision. To address this, we propose OPERA (Objective Perplexity-based Reflective Alignment), which replaces unreliable external judges with intrinsic rewards derived from perplexity dynamics. Specifically, we derive an intrinsic reward signal from perplexity dyna"},"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":"2606.25757","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-06-24T12:31:14Z","cross_cats_sorted":[],"title_canon_sha256":"319c297a5f9b707d72107d68f5dec870ad9024c9422e65367a202667f83682df","abstract_canon_sha256":"8385aae6c152aa4dd031ca8ed93e1cee9be56758f7c80f6628b04996f3148ca0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-25T01:18:14.440220Z","signature_b64":"S0magTbRKnIyILF9JebZIA8tKHRSFo0wvDETD/XCP54tlysM57rPecq4bYsQSgWcQu+z51ByFtuhHUKTrRkbBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5207049ce410f6eb8631cda49eb729050489ccbe2e73b5786089cfefd7e6c764","last_reissued_at":"2026-06-25T01:18:14.439878Z","signature_status":"signed_v1","first_computed_at":"2026-06-25T01:18:14.439878Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"OPERA: Aligning Open-Ended Reasoning via Objective Perplexity-based Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Haoyang Dai, Hongming Tan, Ninghao Liu, Wenxuan Jiang, Xiaoyu Li, Xuecheng Wu, Xuezhi Cao, Zijian Zhang, Zining Fan","submitted_at":"2026-06-24T12:31:14Z","abstract_excerpt":"Reinforcement Learning (RL) has enabled LLMs to excel in objective reasoning tasks such as mathematics and code generation. However, applying RL to open-ended tasks, such as creative writing, remains challenging because LLM-as-a-judge reward models often exhibit stylistic biases and positional inconsistencies, leading to unstable supervision. To address this, we propose OPERA (Objective Perplexity-based Reflective Alignment), which replaces unreliable external judges with intrinsic rewards derived from perplexity dynamics. Specifically, we derive an intrinsic reward signal from perplexity dyna"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.25757","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/2606.25757/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":"2606.25757","created_at":"2026-06-25T01:18:14.439940+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.25757v1","created_at":"2026-06-25T01:18:14.439940+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.25757","created_at":"2026-06-25T01:18:14.439940+00:00"},{"alias_kind":"pith_short_12","alias_value":"KIDQJHHECD3O","created_at":"2026-06-25T01:18:14.439940+00:00"},{"alias_kind":"pith_short_16","alias_value":"KIDQJHHECD3OXBRR","created_at":"2026-06-25T01:18:14.439940+00:00"},{"alias_kind":"pith_short_8","alias_value":"KIDQJHHE","created_at":"2026-06-25T01:18:14.439940+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KIDQJHHECD3OXBRRZWSJ5NZJAU","json":"https://pith.science/pith/KIDQJHHECD3OXBRRZWSJ5NZJAU.json","graph_json":"https://pith.science/api/pith-number/KIDQJHHECD3OXBRRZWSJ5NZJAU/graph.json","events_json":"https://pith.science/api/pith-number/KIDQJHHECD3OXBRRZWSJ5NZJAU/events.json","paper":"https://pith.science/paper/KIDQJHHE"},"agent_actions":{"view_html":"https://pith.science/pith/KIDQJHHECD3OXBRRZWSJ5NZJAU","download_json":"https://pith.science/pith/KIDQJHHECD3OXBRRZWSJ5NZJAU.json","view_paper":"https://pith.science/paper/KIDQJHHE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.25757&json=true","fetch_graph":"https://pith.science/api/pith-number/KIDQJHHECD3OXBRRZWSJ5NZJAU/graph.json","fetch_events":"https://pith.science/api/pith-number/KIDQJHHECD3OXBRRZWSJ5NZJAU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KIDQJHHECD3OXBRRZWSJ5NZJAU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KIDQJHHECD3OXBRRZWSJ5NZJAU/action/storage_attestation","attest_author":"https://pith.science/pith/KIDQJHHECD3OXBRRZWSJ5NZJAU/action/author_attestation","sign_citation":"https://pith.science/pith/KIDQJHHECD3OXBRRZWSJ5NZJAU/action/citation_signature","submit_replication":"https://pith.science/pith/KIDQJHHECD3OXBRRZWSJ5NZJAU/action/replication_record"}},"created_at":"2026-06-25T01:18:14.439940+00:00","updated_at":"2026-06-25T01:18:14.439940+00:00"}