{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:6KAY4W3HOISRER77RS6GJE7OUU","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":"afa466e83cf3cac1ee7a8636d154963c1fd22511518eff960fbf48953152f33f","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T20:57:57Z","title_canon_sha256":"6a680eaa346257d11b2a6706e7d81efdad50682bdb862bbcd818e490000e4c01"},"schema_version":"1.0","source":{"id":"2606.05434","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.05434","created_at":"2026-06-05T00:13:57Z"},{"alias_kind":"arxiv_version","alias_value":"2606.05434v1","created_at":"2026-06-05T00:13:57Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.05434","created_at":"2026-06-05T00:13:57Z"},{"alias_kind":"pith_short_12","alias_value":"6KAY4W3HOISR","created_at":"2026-06-05T00:13:57Z"},{"alias_kind":"pith_short_16","alias_value":"6KAY4W3HOISRER77","created_at":"2026-06-05T00:13:57Z"},{"alias_kind":"pith_short_8","alias_value":"6KAY4W3H","created_at":"2026-06-05T00:13:57Z"}],"graph_snapshots":[{"event_id":"sha256:f1d3b2d4570e1e0fb29b0bbf4f104046a8ab58c01c51fbba39b40a11fa8eb70b","target":"graph","created_at":"2026-06-05T00:13:57Z","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/2606.05434/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Group Relative Policy Optimisation (GRPO) has emerged as an effective reinforcement-learning algorithm for aligning language models on reasoning tasks, but it treats every token position and every sampled rollout symmetrically. We introduce two complementary extensions: (i) Adaptive-Horizon GRPO (AH-GRPO), which weights each token's policy gradient using a cumulative entropy-based discount that reduces the effective horizon when the model is uncertain, and (ii) Selective-Advantage AH-GRPO (SA-AH-GRPO), which applies this discounting only to negative-advantage rollouts, leaving positive-advanta","authors_text":"Chirag Chawla, Madhav S. Baidya, Rohan Charudatt Salvi","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T20:57:57Z","title":"Selective-Advantage Entropy-Adaptive Horizon GRPO: Asymmetric Token-Level Discounting for Efficient Reinforcement Learning of Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.05434","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":"sha256:294311f78325ac577c12142b71cb0ae4ece7f5caafca6e747e0dea5b8a420d50","target":"record","created_at":"2026-06-05T00:13:57Z","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":"afa466e83cf3cac1ee7a8636d154963c1fd22511518eff960fbf48953152f33f","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-03T20:57:57Z","title_canon_sha256":"6a680eaa346257d11b2a6706e7d81efdad50682bdb862bbcd818e490000e4c01"},"schema_version":"1.0","source":{"id":"2606.05434","kind":"arxiv","version":1}},"canonical_sha256":"f2818e5b6772251247ff8cbc6493eea506ef7ed507afe553d84c316a1c5ad3d3","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f2818e5b6772251247ff8cbc6493eea506ef7ed507afe553d84c316a1c5ad3d3","first_computed_at":"2026-06-05T00:13:57.645337Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-05T00:13:57.645337Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"6MqdghlG4PMXNV2vC9WV8lzrnP8PcLH8srHOk0bz/G1aGwrwu6gVIGMLKddgCvXrbdUwBa3M/25rYUZhAtDBCg==","signature_status":"signed_v1","signed_at":"2026-06-05T00:13:57.645992Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.05434","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:294311f78325ac577c12142b71cb0ae4ece7f5caafca6e747e0dea5b8a420d50","sha256:f1d3b2d4570e1e0fb29b0bbf4f104046a8ab58c01c51fbba39b40a11fa8eb70b"],"state_sha256":"d7692bb28817f81528e4a3c3076441c167cb446ba2cab62113696d476b3e1259"}