{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:PPFRUR3YKO2ZRSS262TGN5K7SG","short_pith_number":"pith:PPFRUR3Y","schema_version":"1.0","canonical_sha256":"7bcb1a477853b598ca5af6a666f55f91a48a0c24ad4f10b6f34f241cd313f9b9","source":{"kind":"arxiv","id":"2407.15902","version":2},"attestation_state":"computed","paper":{"title":"Revisiting the Robust Alignment of Circuit Breakers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Leo Schwinn, Simon Geisler","submitted_at":"2024-07-22T11:53:48Z","abstract_excerpt":"Over the past decade, adversarial training has emerged as one of the few reliable methods for enhancing model robustness against adversarial attacks [Szegedy et al., 2014, Madry et al., 2018, Xhonneux et al., 2024], while many alternative approaches have failed to withstand rigorous subsequent evaluations. Recently, an alternative defense mechanism, namely \"circuit breakers\" [Zou et al., 2024], has shown promising results for aligning LLMs. In this report, we show that the robustness claims of \"Improving Alignment and Robustness with Circuit Breakers\" against unconstraint continuous attacks in"},"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":"2407.15902","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2024-07-22T11:53:48Z","cross_cats_sorted":[],"title_canon_sha256":"99a4fbae933c9f9535da37e922b77d8a5eaae116f859e533c1e7a06847100661","abstract_canon_sha256":"777d790cd6494a80964b8bfd0cb1e12487b214cca7352dd42eef904051423fec"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:51:18.862409Z","signature_b64":"r/FQVNdhK/2xonVxeHmaT1mhl/6Y2iCy/hhGDXDufmW17MGwC3maTap9Jv52WFo4kFMw8MVnzwbhAfOmc5fiBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7bcb1a477853b598ca5af6a666f55f91a48a0c24ad4f10b6f34f241cd313f9b9","last_reissued_at":"2026-07-05T08:51:18.861936Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:51:18.861936Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Revisiting the Robust Alignment of Circuit Breakers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CR","authors_text":"Leo Schwinn, Simon Geisler","submitted_at":"2024-07-22T11:53:48Z","abstract_excerpt":"Over the past decade, adversarial training has emerged as one of the few reliable methods for enhancing model robustness against adversarial attacks [Szegedy et al., 2014, Madry et al., 2018, Xhonneux et al., 2024], while many alternative approaches have failed to withstand rigorous subsequent evaluations. Recently, an alternative defense mechanism, namely \"circuit breakers\" [Zou et al., 2024], has shown promising results for aligning LLMs. In this report, we show that the robustness claims of \"Improving Alignment and Robustness with Circuit Breakers\" against unconstraint continuous attacks in"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.15902","kind":"arxiv","version":2},"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/2407.15902/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":"2407.15902","created_at":"2026-07-05T08:51:18.861992+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.15902v2","created_at":"2026-07-05T08:51:18.861992+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.15902","created_at":"2026-07-05T08:51:18.861992+00:00"},{"alias_kind":"pith_short_12","alias_value":"PPFRUR3YKO2Z","created_at":"2026-07-05T08:51:18.861992+00:00"},{"alias_kind":"pith_short_16","alias_value":"PPFRUR3YKO2ZRSS2","created_at":"2026-07-05T08:51:18.861992+00:00"},{"alias_kind":"pith_short_8","alias_value":"PPFRUR3Y","created_at":"2026-07-05T08:51:18.861992+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.19321","citing_title":"Exploring and Developing a Pre-Model Safeguard with Draft Models","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03095","citing_title":"Revisiting JBShield: Breaking and Rebuilding Representation-Level Jailbreak Defenses","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09544","citing_title":"Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm Types","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09544","citing_title":"Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm Types","ref_index":22,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PPFRUR3YKO2ZRSS262TGN5K7SG","json":"https://pith.science/pith/PPFRUR3YKO2ZRSS262TGN5K7SG.json","graph_json":"https://pith.science/api/pith-number/PPFRUR3YKO2ZRSS262TGN5K7SG/graph.json","events_json":"https://pith.science/api/pith-number/PPFRUR3YKO2ZRSS262TGN5K7SG/events.json","paper":"https://pith.science/paper/PPFRUR3Y"},"agent_actions":{"view_html":"https://pith.science/pith/PPFRUR3YKO2ZRSS262TGN5K7SG","download_json":"https://pith.science/pith/PPFRUR3YKO2ZRSS262TGN5K7SG.json","view_paper":"https://pith.science/paper/PPFRUR3Y","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.15902&json=true","fetch_graph":"https://pith.science/api/pith-number/PPFRUR3YKO2ZRSS262TGN5K7SG/graph.json","fetch_events":"https://pith.science/api/pith-number/PPFRUR3YKO2ZRSS262TGN5K7SG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PPFRUR3YKO2ZRSS262TGN5K7SG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PPFRUR3YKO2ZRSS262TGN5K7SG/action/storage_attestation","attest_author":"https://pith.science/pith/PPFRUR3YKO2ZRSS262TGN5K7SG/action/author_attestation","sign_citation":"https://pith.science/pith/PPFRUR3YKO2ZRSS262TGN5K7SG/action/citation_signature","submit_replication":"https://pith.science/pith/PPFRUR3YKO2ZRSS262TGN5K7SG/action/replication_record"}},"created_at":"2026-07-05T08:51:18.861992+00:00","updated_at":"2026-07-05T08:51:18.861992+00:00"}