{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:ADYDZCI7JFSRIUXWZP2YCQQ54J","short_pith_number":"pith:ADYDZCI7","schema_version":"1.0","canonical_sha256":"00f03c891f49651452f6cbf581421de27d286f147e4462c2457c10aaa1c180aa","source":{"kind":"arxiv","id":"2606.17257","version":1},"attestation_state":"computed","paper":{"title":"Pulling The REINS: Training-Free Safety Alignment of Video Diffusion Models via Representation Steering","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Amit K. Roy-Chowdhury, Arindam Dutta, Athula Balachandran, Rohit Kundu, Sarosij Bose","submitted_at":"2026-06-15T20:03:09Z","abstract_excerpt":"Open-weight video diffusion models can generate photorealistic unsafe content, from violence to misinformation, yet existing defenses either require expensive safety fine-tuning that degrades general capability, or apply external filters that are trivially bypassed by adversarial prompts. We present REINS (REpresentation-space INference-time Safety steering), a training-free method that aligns video diffusion models at inference time by steering their internal representations toward safe generation. Our key finding is that safety-relevant structure is linearly encoded in the hidden-state activ"},"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":"2606.17257","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-06-15T20:03:09Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b4c5e89413a4fb23bc16ac1692c21d4eb27edf6ce24db8f53b24be989da5f6ab","abstract_canon_sha256":"ae625945a1dbce06236810783ad8a7905b4d5174b4eb046a62544c139979d3b7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-19T16:10:07.148988Z","signature_b64":"OPPR8EuN1rG+4VaLKFalJ9q7fbN7dOa/o4fMTqCsQSMl9SVcfbkdiNUX3yq/rUoI4pP2ZvgoFFs/HamIs3glAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"00f03c891f49651452f6cbf581421de27d286f147e4462c2457c10aaa1c180aa","last_reissued_at":"2026-06-19T16:10:07.148659Z","signature_status":"signed_v1","first_computed_at":"2026-06-19T16:10:07.148659Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Pulling The REINS: Training-Free Safety Alignment of Video Diffusion Models via Representation Steering","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Amit K. Roy-Chowdhury, Arindam Dutta, Athula Balachandran, Rohit Kundu, Sarosij Bose","submitted_at":"2026-06-15T20:03:09Z","abstract_excerpt":"Open-weight video diffusion models can generate photorealistic unsafe content, from violence to misinformation, yet existing defenses either require expensive safety fine-tuning that degrades general capability, or apply external filters that are trivially bypassed by adversarial prompts. We present REINS (REpresentation-space INference-time Safety steering), a training-free method that aligns video diffusion models at inference time by steering their internal representations toward safe generation. Our key finding is that safety-relevant structure is linearly encoded in the hidden-state activ"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.17257","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/2606.17257/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":"2606.17257","created_at":"2026-06-19T16:10:07.148711+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.17257v1","created_at":"2026-06-19T16:10:07.148711+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.17257","created_at":"2026-06-19T16:10:07.148711+00:00"},{"alias_kind":"pith_short_12","alias_value":"ADYDZCI7JFSR","created_at":"2026-06-19T16:10:07.148711+00:00"},{"alias_kind":"pith_short_16","alias_value":"ADYDZCI7JFSRIUXW","created_at":"2026-06-19T16:10:07.148711+00:00"},{"alias_kind":"pith_short_8","alias_value":"ADYDZCI7","created_at":"2026-06-19T16:10:07.148711+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/ADYDZCI7JFSRIUXWZP2YCQQ54J","json":"https://pith.science/pith/ADYDZCI7JFSRIUXWZP2YCQQ54J.json","graph_json":"https://pith.science/api/pith-number/ADYDZCI7JFSRIUXWZP2YCQQ54J/graph.json","events_json":"https://pith.science/api/pith-number/ADYDZCI7JFSRIUXWZP2YCQQ54J/events.json","paper":"https://pith.science/paper/ADYDZCI7"},"agent_actions":{"view_html":"https://pith.science/pith/ADYDZCI7JFSRIUXWZP2YCQQ54J","download_json":"https://pith.science/pith/ADYDZCI7JFSRIUXWZP2YCQQ54J.json","view_paper":"https://pith.science/paper/ADYDZCI7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.17257&json=true","fetch_graph":"https://pith.science/api/pith-number/ADYDZCI7JFSRIUXWZP2YCQQ54J/graph.json","fetch_events":"https://pith.science/api/pith-number/ADYDZCI7JFSRIUXWZP2YCQQ54J/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ADYDZCI7JFSRIUXWZP2YCQQ54J/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ADYDZCI7JFSRIUXWZP2YCQQ54J/action/storage_attestation","attest_author":"https://pith.science/pith/ADYDZCI7JFSRIUXWZP2YCQQ54J/action/author_attestation","sign_citation":"https://pith.science/pith/ADYDZCI7JFSRIUXWZP2YCQQ54J/action/citation_signature","submit_replication":"https://pith.science/pith/ADYDZCI7JFSRIUXWZP2YCQQ54J/action/replication_record"}},"created_at":"2026-06-19T16:10:07.148711+00:00","updated_at":"2026-06-19T16:10:07.148711+00:00"}