{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:QWE7CP4JZX375DTIZNE5MWROJG","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":"92be38605f9d6f4c30bd09e5a3efa4ac33d0a1399e56ee855e6509c8282ad8b6","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-18T17:50:04Z","title_canon_sha256":"688ccdd1eb5de41375c1154406b5d8f3e9b21e68b937c7dc772c25a65fa96d30"},"schema_version":"1.0","source":{"id":"2605.18719","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.18719","created_at":"2026-05-20T00:06:16Z"},{"alias_kind":"arxiv_version","alias_value":"2605.18719v1","created_at":"2026-05-20T00:06:16Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.18719","created_at":"2026-05-20T00:06:16Z"},{"alias_kind":"pith_short_12","alias_value":"QWE7CP4JZX37","created_at":"2026-05-20T00:06:16Z"},{"alias_kind":"pith_short_16","alias_value":"QWE7CP4JZX375DTI","created_at":"2026-05-20T00:06:16Z"},{"alias_kind":"pith_short_8","alias_value":"QWE7CP4J","created_at":"2026-05-20T00:06:16Z"}],"graph_snapshots":[{"event_id":"sha256:99d28e9538c19d4ae1a046e30c295d528b21feef34737dae8b9b3706922a31f1","target":"graph","created_at":"2026-05-20T00:06:16Z","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":[{"findings_count":0,"name":"claim_evidence","ran_at":"2026-05-20T00:01:59.043508Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2605.18719/integrity.json","findings":[],"snapshot_sha256":"629a37b62915818045a3708bd2df71c4857eca19b790a66ce153ae022e43cd85","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Diffusion models have been widely studied for removing unsafe content learned during pre-training. Existing methods require expensive supervised data, either unsafe-text paired with safe-image groundtruth or negative/positive image pairs, making them impractical to scale. Furthermore, offline reinforcement learning and supervised fine-tuning approaches that generate synthetic data offline suffer from catastrophic forgetting, degrading generation quality. We propose a novel online reinforcement learning framework that addresses both data scarcity and model degradation through post-training with","authors_text":"Abhishek Basu, Ankan Deria, Fahad Shamshad, Hisham Cholakkal, Karthik Nandakumar, Komal Kumar","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-18T17:50:04Z","title":"SafeDiffusion-R1: Online Reward Steering for Safe Diffusion Post-Training"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.18719","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:a464e7da65a410805717c3903c60fbe1189f836313fbe772b8cc5235f2a25206","target":"record","created_at":"2026-05-20T00:06:16Z","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":"92be38605f9d6f4c30bd09e5a3efa4ac33d0a1399e56ee855e6509c8282ad8b6","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-18T17:50:04Z","title_canon_sha256":"688ccdd1eb5de41375c1154406b5d8f3e9b21e68b937c7dc772c25a65fa96d30"},"schema_version":"1.0","source":{"id":"2605.18719","kind":"arxiv","version":1}},"canonical_sha256":"8589f13f89cdf7fe8e68cb49d65a2e4991bc54d9b6f2dcdf5b5579c1a13e2c68","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8589f13f89cdf7fe8e68cb49d65a2e4991bc54d9b6f2dcdf5b5579c1a13e2c68","first_computed_at":"2026-05-20T00:06:16.825284Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:06:16.825284Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"RlFLn/AnN5VLvPdQlFkW3mPWnJZp8vIdzXe2vXC/dm99HSdEui9eSJLkckVZw3CPTvJxRNTO5AJBDMyOuiAUCw==","signature_status":"signed_v1","signed_at":"2026-05-20T00:06:16.826136Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.18719","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a464e7da65a410805717c3903c60fbe1189f836313fbe772b8cc5235f2a25206","sha256:99d28e9538c19d4ae1a046e30c295d528b21feef34737dae8b9b3706922a31f1"],"state_sha256":"2de26755dd89401d050320e361c04edb230c3fc125d5d6ce1aa64cee96fd8136"}