{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:BMJZILZGGQXKQW6FFANSYLUEON","short_pith_number":"pith:BMJZILZG","canonical_record":{"source":{"id":"2605.29009","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T19:04:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e0a9afa4588bc810c8c806ea92d252d64aaf0ec924af7e82eacd025aa4fac92d","abstract_canon_sha256":"159903be559604308163bc3a0711dcc8017583584bcd35cc0b16d68a8b4b5ec7"},"schema_version":"1.0"},"canonical_sha256":"0b13942f26342ea85bc5281b2c2e8473604ccfbd1ff4b60291c4b045cede4e89","source":{"kind":"arxiv","id":"2605.29009","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.29009","created_at":"2026-05-29T01:04:43Z"},{"alias_kind":"arxiv_version","alias_value":"2605.29009v1","created_at":"2026-05-29T01:04:43Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.29009","created_at":"2026-05-29T01:04:43Z"},{"alias_kind":"pith_short_12","alias_value":"BMJZILZGGQXK","created_at":"2026-05-29T01:04:43Z"},{"alias_kind":"pith_short_16","alias_value":"BMJZILZGGQXKQW6F","created_at":"2026-05-29T01:04:43Z"},{"alias_kind":"pith_short_8","alias_value":"BMJZILZG","created_at":"2026-05-29T01:04:43Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:BMJZILZGGQXKQW6FFANSYLUEON","target":"record","payload":{"canonical_record":{"source":{"id":"2605.29009","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T19:04:35Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"e0a9afa4588bc810c8c806ea92d252d64aaf0ec924af7e82eacd025aa4fac92d","abstract_canon_sha256":"159903be559604308163bc3a0711dcc8017583584bcd35cc0b16d68a8b4b5ec7"},"schema_version":"1.0"},"canonical_sha256":"0b13942f26342ea85bc5281b2c2e8473604ccfbd1ff4b60291c4b045cede4e89","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-29T01:04:43.370185Z","signature_b64":"zrsknw+d2cRuu/1CKFY+oS8DYJ5p8JfgdCRfY7fKwwavFwf1A9j93DctrxEy2na6s0tiCNwLKl/SCb2AQRJIDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0b13942f26342ea85bc5281b2c2e8473604ccfbd1ff4b60291c4b045cede4e89","last_reissued_at":"2026-05-29T01:04:43.369788Z","signature_status":"signed_v1","first_computed_at":"2026-05-29T01:04:43.369788Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.29009","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":"2026-05-29T01:04:43Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+ht3HIs6mgQ49W8MkydhZkxXetL2rwynHHnSpspC4YMtDiKCtioCkjp7BxJgwmOZFIMxpHHGHd1nFQ2xeUYUCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-22T23:27:24.347646Z"},"content_sha256":"e3ed10e342760bb5a77b98393cd86b974ef30775e580f832659facf48f1624d5","schema_version":"1.0","event_id":"sha256:e3ed10e342760bb5a77b98393cd86b974ef30775e580f832659facf48f1624d5"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:BMJZILZGGQXKQW6FFANSYLUEON","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Label-Free Reinforcement Learning via Cross-Model Entropy","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Hossein Shirazi, Matt Gorbett","submitted_at":"2026-05-27T19:04:35Z","abstract_excerpt":"Post-training large language models with reinforcement learning is bottlenecked by the reward signal. Existing approaches require either ground-truth verifiable rewards, restricting training to domains with automatic correctness checks (e.g., mathematics, code execution), or human preference labels, which are expensive to collect and prone to reward hacking. Recent label-free methods replace ground-truth verifiers with self-referential signals like majority voting or token entropy over a model's own outputs, but risk reinforcing a model's own errors. In this work we propose Cross-Model Entropy"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.29009","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/2605.29009/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":"2026-05-29T01:04:43Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CZroBySa9VA0ybn1fRH0EPJpHFThsYdPo37JNz4LYKBWB3Iari0AfBeN0goFLEMrY7WPteooskljIW8e5gkSCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-22T23:27:24.348020Z"},"content_sha256":"d889f2543e42f0cdf68026f130622b60a87438929cc67a8048aca0b8f4880654","schema_version":"1.0","event_id":"sha256:d889f2543e42f0cdf68026f130622b60a87438929cc67a8048aca0b8f4880654"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BMJZILZGGQXKQW6FFANSYLUEON/bundle.json","state_url":"https://pith.science/pith/BMJZILZGGQXKQW6FFANSYLUEON/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BMJZILZGGQXKQW6FFANSYLUEON/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":"2026-06-22T23:27:24Z","links":{"resolver":"https://pith.science/pith/BMJZILZGGQXKQW6FFANSYLUEON","bundle":"https://pith.science/pith/BMJZILZGGQXKQW6FFANSYLUEON/bundle.json","state":"https://pith.science/pith/BMJZILZGGQXKQW6FFANSYLUEON/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BMJZILZGGQXKQW6FFANSYLUEON/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:BMJZILZGGQXKQW6FFANSYLUEON","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":"159903be559604308163bc3a0711dcc8017583584bcd35cc0b16d68a8b4b5ec7","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T19:04:35Z","title_canon_sha256":"e0a9afa4588bc810c8c806ea92d252d64aaf0ec924af7e82eacd025aa4fac92d"},"schema_version":"1.0","source":{"id":"2605.29009","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.29009","created_at":"2026-05-29T01:04:43Z"},{"alias_kind":"arxiv_version","alias_value":"2605.29009v1","created_at":"2026-05-29T01:04:43Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.29009","created_at":"2026-05-29T01:04:43Z"},{"alias_kind":"pith_short_12","alias_value":"BMJZILZGGQXK","created_at":"2026-05-29T01:04:43Z"},{"alias_kind":"pith_short_16","alias_value":"BMJZILZGGQXKQW6F","created_at":"2026-05-29T01:04:43Z"},{"alias_kind":"pith_short_8","alias_value":"BMJZILZG","created_at":"2026-05-29T01:04:43Z"}],"graph_snapshots":[{"event_id":"sha256:d889f2543e42f0cdf68026f130622b60a87438929cc67a8048aca0b8f4880654","target":"graph","created_at":"2026-05-29T01:04:43Z","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/2605.29009/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Post-training large language models with reinforcement learning is bottlenecked by the reward signal. Existing approaches require either ground-truth verifiable rewards, restricting training to domains with automatic correctness checks (e.g., mathematics, code execution), or human preference labels, which are expensive to collect and prone to reward hacking. Recent label-free methods replace ground-truth verifiers with self-referential signals like majority voting or token entropy over a model's own outputs, but risk reinforcing a model's own errors. In this work we propose Cross-Model Entropy","authors_text":"Hossein Shirazi, Matt Gorbett","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T19:04:35Z","title":"Label-Free Reinforcement Learning via Cross-Model Entropy"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.29009","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:e3ed10e342760bb5a77b98393cd86b974ef30775e580f832659facf48f1624d5","target":"record","created_at":"2026-05-29T01:04:43Z","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":"159903be559604308163bc3a0711dcc8017583584bcd35cc0b16d68a8b4b5ec7","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-27T19:04:35Z","title_canon_sha256":"e0a9afa4588bc810c8c806ea92d252d64aaf0ec924af7e82eacd025aa4fac92d"},"schema_version":"1.0","source":{"id":"2605.29009","kind":"arxiv","version":1}},"canonical_sha256":"0b13942f26342ea85bc5281b2c2e8473604ccfbd1ff4b60291c4b045cede4e89","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0b13942f26342ea85bc5281b2c2e8473604ccfbd1ff4b60291c4b045cede4e89","first_computed_at":"2026-05-29T01:04:43.369788Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-29T01:04:43.369788Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"zrsknw+d2cRuu/1CKFY+oS8DYJ5p8JfgdCRfY7fKwwavFwf1A9j93DctrxEy2na6s0tiCNwLKl/SCb2AQRJIDA==","signature_status":"signed_v1","signed_at":"2026-05-29T01:04:43.370185Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.29009","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e3ed10e342760bb5a77b98393cd86b974ef30775e580f832659facf48f1624d5","sha256:d889f2543e42f0cdf68026f130622b60a87438929cc67a8048aca0b8f4880654"],"state_sha256":"111c27fc0fecf132bc0af86b5596eda2af20085927e7bf9f72fc1e439d1314d7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HXyDCRHKWT4MTW1MDvT1piJQgw+mZ6tnJhHXlVn0sTwfgXqACh9hFB3gK27IBVpLNv3Us6N3uRXZnVp0kuzMCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-22T23:27:24.350034Z","bundle_sha256":"83251b60c7bf8d2996cd23a44d48db072fead2c8974f8d27d3a461f8fbcda4f8"}}