{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:WLRYMFOBNVG2IKN6LVEMEIL6KE","short_pith_number":"pith:WLRYMFOB","canonical_record":{"source":{"id":"2605.22620","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T15:30:47Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"315a4f1c39141160a157b1e00506e84a582c98fbfbe969f501619455875d6459","abstract_canon_sha256":"9943407b710753cf41415796399438d4572a3c7d7dbffdc89b5b299773dade7e"},"schema_version":"1.0"},"canonical_sha256":"b2e38615c16d4da429be5d48c2217e512ef42886a2dcccbce79ef319966dd8ea","source":{"kind":"arxiv","id":"2605.22620","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.22620","created_at":"2026-05-22T01:05:00Z"},{"alias_kind":"arxiv_version","alias_value":"2605.22620v1","created_at":"2026-05-22T01:05:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.22620","created_at":"2026-05-22T01:05:00Z"},{"alias_kind":"pith_short_12","alias_value":"WLRYMFOBNVG2","created_at":"2026-05-22T01:05:00Z"},{"alias_kind":"pith_short_16","alias_value":"WLRYMFOBNVG2IKN6","created_at":"2026-05-22T01:05:00Z"},{"alias_kind":"pith_short_8","alias_value":"WLRYMFOB","created_at":"2026-05-22T01:05:00Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:WLRYMFOBNVG2IKN6LVEMEIL6KE","target":"record","payload":{"canonical_record":{"source":{"id":"2605.22620","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T15:30:47Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"315a4f1c39141160a157b1e00506e84a582c98fbfbe969f501619455875d6459","abstract_canon_sha256":"9943407b710753cf41415796399438d4572a3c7d7dbffdc89b5b299773dade7e"},"schema_version":"1.0"},"canonical_sha256":"b2e38615c16d4da429be5d48c2217e512ef42886a2dcccbce79ef319966dd8ea","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-22T01:05:00.877615Z","signature_b64":"MXx996Nto7U4q+VaUt13kdlitMmK11ENskRubqsDan6F7SO1up2bBeFWWb5cn3BAe1UX8tTMQoO1F8ePDES0AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b2e38615c16d4da429be5d48c2217e512ef42886a2dcccbce79ef319966dd8ea","last_reissued_at":"2026-05-22T01:05:00.877020Z","signature_status":"signed_v1","first_computed_at":"2026-05-22T01:05:00.877020Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.22620","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":"2026-05-22T01:05:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"JOv4h3OYoWi171vJU64F5kfwWZ8AM7uTYW6pYS/c93yownEm3B7BKFfzk7bauM6B7uusob+B+R7fEPC5XObNBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-02T02:04:50.137279Z"},"content_sha256":"3314090c0c2abb1f17e7462761bbb28b24813779f97baa24038246683c9fe6ff","schema_version":"1.0","event_id":"sha256:3314090c0c2abb1f17e7462761bbb28b24813779f97baa24038246683c9fe6ff"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:WLRYMFOBNVG2IKN6LVEMEIL6KE","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Two is better than one: A Collapse-free Multi-Reward RLIF Training Framework","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Ahsan Habib Akash, Binod Bhattarai, Diganta Sikdar, Prashnna Gyawali, Shourov Joarder","submitted_at":"2026-05-21T15:30:47Z","abstract_excerpt":"Reinforcement learning with verifiable rewards (RLVR) has substantially improved the reasoning ability of LLMs, but often depends on external supervision from human annotations or gold-standard solutions. Reinforcement learning from internal feedback (RLIF) has recently emerged as a scalable unsupervised alternative, using signals extracted from the model itself. However, existing RLIF methods typically rely on a single internal reward, which can lead to reward hacking, entropy collapse, and degraded reasoning structure. We propose a multi-reward RLIF framework that decomposes the training sig"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.22620","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/2605.22620/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":"2026-05-22T01:05:00Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Iktp6zUXjI0NP+g2eZwm/jKwQESlNBYNC1Wo2NWYtBbLh7p+yo4c15mqOmsIs4P6SkHMkd3PD2oy55miOTVwAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-02T02:04:50.137672Z"},"content_sha256":"116b1ec153f43671e76a712e110115e6164125a66021a6eef9a23e7a43b61cdc","schema_version":"1.0","event_id":"sha256:116b1ec153f43671e76a712e110115e6164125a66021a6eef9a23e7a43b61cdc"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/WLRYMFOBNVG2IKN6LVEMEIL6KE/bundle.json","state_url":"https://pith.science/pith/WLRYMFOBNVG2IKN6LVEMEIL6KE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/WLRYMFOBNVG2IKN6LVEMEIL6KE/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":"2026-06-02T02:04:50Z","links":{"resolver":"https://pith.science/pith/WLRYMFOBNVG2IKN6LVEMEIL6KE","bundle":"https://pith.science/pith/WLRYMFOBNVG2IKN6LVEMEIL6KE/bundle.json","state":"https://pith.science/pith/WLRYMFOBNVG2IKN6LVEMEIL6KE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/WLRYMFOBNVG2IKN6LVEMEIL6KE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:WLRYMFOBNVG2IKN6LVEMEIL6KE","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":"9943407b710753cf41415796399438d4572a3c7d7dbffdc89b5b299773dade7e","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T15:30:47Z","title_canon_sha256":"315a4f1c39141160a157b1e00506e84a582c98fbfbe969f501619455875d6459"},"schema_version":"1.0","source":{"id":"2605.22620","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.22620","created_at":"2026-05-22T01:05:00Z"},{"alias_kind":"arxiv_version","alias_value":"2605.22620v1","created_at":"2026-05-22T01:05:00Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.22620","created_at":"2026-05-22T01:05:00Z"},{"alias_kind":"pith_short_12","alias_value":"WLRYMFOBNVG2","created_at":"2026-05-22T01:05:00Z"},{"alias_kind":"pith_short_16","alias_value":"WLRYMFOBNVG2IKN6","created_at":"2026-05-22T01:05:00Z"},{"alias_kind":"pith_short_8","alias_value":"WLRYMFOB","created_at":"2026-05-22T01:05:00Z"}],"graph_snapshots":[{"event_id":"sha256:116b1ec153f43671e76a712e110115e6164125a66021a6eef9a23e7a43b61cdc","target":"graph","created_at":"2026-05-22T01:05:00Z","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/2605.22620/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Reinforcement learning with verifiable rewards (RLVR) has substantially improved the reasoning ability of LLMs, but often depends on external supervision from human annotations or gold-standard solutions. Reinforcement learning from internal feedback (RLIF) has recently emerged as a scalable unsupervised alternative, using signals extracted from the model itself. However, existing RLIF methods typically rely on a single internal reward, which can lead to reward hacking, entropy collapse, and degraded reasoning structure. We propose a multi-reward RLIF framework that decomposes the training sig","authors_text":"Ahsan Habib Akash, Binod Bhattarai, Diganta Sikdar, Prashnna Gyawali, Shourov Joarder","cross_cats":["cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T15:30:47Z","title":"Two is better than one: A Collapse-free Multi-Reward RLIF Training Framework"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.22620","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":"sha256:3314090c0c2abb1f17e7462761bbb28b24813779f97baa24038246683c9fe6ff","target":"record","created_at":"2026-05-22T01:05:00Z","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":"9943407b710753cf41415796399438d4572a3c7d7dbffdc89b5b299773dade7e","cross_cats_sorted":["cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-21T15:30:47Z","title_canon_sha256":"315a4f1c39141160a157b1e00506e84a582c98fbfbe969f501619455875d6459"},"schema_version":"1.0","source":{"id":"2605.22620","kind":"arxiv","version":1}},"canonical_sha256":"b2e38615c16d4da429be5d48c2217e512ef42886a2dcccbce79ef319966dd8ea","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"b2e38615c16d4da429be5d48c2217e512ef42886a2dcccbce79ef319966dd8ea","first_computed_at":"2026-05-22T01:05:00.877020Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-22T01:05:00.877020Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"MXx996Nto7U4q+VaUt13kdlitMmK11ENskRubqsDan6F7SO1up2bBeFWWb5cn3BAe1UX8tTMQoO1F8ePDES0AA==","signature_status":"signed_v1","signed_at":"2026-05-22T01:05:00.877615Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.22620","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3314090c0c2abb1f17e7462761bbb28b24813779f97baa24038246683c9fe6ff","sha256:116b1ec153f43671e76a712e110115e6164125a66021a6eef9a23e7a43b61cdc"],"state_sha256":"2d9a3bf203e4aae298d51ed9bee308cdf40ca8ec6bbfad72cab8a4beae0b29d9"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CprvGEKGiAFIa44R8KZhIc60YKtZ8ctb+x/tapYIACOGpg/Ve/l0l2tDmVEH4iTfrtjWoPXNMq1J8Am+WC09BA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-02T02:04:50.139712Z","bundle_sha256":"483d4029bc99732ecd03b53717418dd67634b62336bd8df6d844470353bf403f"}}