{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:FQ5LFIOMJGGG54HJNL567P72SG","short_pith_number":"pith:FQ5LFIOM","canonical_record":{"source":{"id":"2506.05451","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2025-06-05T17:56:05Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"dd8e9f89d97cb1d95ca6225cc1ef07b22d51c7658e2c2506c155b9571da72c7e","abstract_canon_sha256":"3e991ade848579be124829cf6a58359c632238ed6fdcc86aebe639ee1bff998b"},"schema_version":"1.0"},"canonical_sha256":"2c3ab2a1cc498c6ef0e96afbefbffa91a8627b4c04bfab99bd9f69e8e261fdb4","source":{"kind":"arxiv","id":"2506.05451","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.05451","created_at":"2026-07-05T11:17:02Z"},{"alias_kind":"arxiv_version","alias_value":"2506.05451v1","created_at":"2026-07-05T11:17:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.05451","created_at":"2026-07-05T11:17:02Z"},{"alias_kind":"pith_short_12","alias_value":"FQ5LFIOMJGGG","created_at":"2026-07-05T11:17:02Z"},{"alias_kind":"pith_short_16","alias_value":"FQ5LFIOMJGGG54HJ","created_at":"2026-07-05T11:17:02Z"},{"alias_kind":"pith_short_8","alias_value":"FQ5LFIOM","created_at":"2026-07-05T11:17:02Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:FQ5LFIOMJGGG54HJNL567P72SG","target":"record","payload":{"canonical_record":{"source":{"id":"2506.05451","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2025-06-05T17:56:05Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"dd8e9f89d97cb1d95ca6225cc1ef07b22d51c7658e2c2506c155b9571da72c7e","abstract_canon_sha256":"3e991ade848579be124829cf6a58359c632238ed6fdcc86aebe639ee1bff998b"},"schema_version":"1.0"},"canonical_sha256":"2c3ab2a1cc498c6ef0e96afbefbffa91a8627b4c04bfab99bd9f69e8e261fdb4","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:17:02.991005Z","signature_b64":"7QaAXFYo19rgIoc0JjchmIFjYcEaQFNsCKg5xBKYW6EGsCBr4MtNxKbFTYX5BmSmIFQ120QenH+vZgn5bhQBDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2c3ab2a1cc498c6ef0e96afbefbffa91a8627b4c04bfab99bd9f69e8e261fdb4","last_reissued_at":"2026-07-05T11:17:02.990526Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:17:02.990526Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2506.05451","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":"2026-07-05T11:17:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kwdflVFMkOVuq7QnOEsdPBsDgq2nchCR4JxY2JnG0sNroLzpyEuW82y6rZDxpbYO9K/SSj9b4kGImSdB6CsNDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T03:27:47.910062Z"},"content_sha256":"b05255df23ab4e5d15ae25e451e3475fd9e88066f5eb47aa9201aa5942af3ae6","schema_version":"1.0","event_id":"sha256:b05255df23ab4e5d15ae25e451e3475fd9e88066f5eb47aa9201aa5942af3ae6"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:FQ5LFIOMJGGG54HJNL567P72SG","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.SE","authors_text":"Aeree Cho, Duen Horng Chau, Grace C. Kim, Mansi Phute, Seongmin Lee, ShengYun Peng","submitted_at":"2025-06-05T17:56:05Z","abstract_excerpt":"As large language models (LLMs) see wider real-world use, understanding and mitigating their unsafe behaviors is critical. Interpretation techniques can reveal causes of unsafe outputs and guide safety, but such connections with safety are often overlooked in prior surveys. We present the first survey that bridges this gap, introducing a unified framework that connects safety-focused interpretation methods, the safety enhancements they inform, and the tools that operationalize them. Our novel taxonomy, organized by LLM workflow stages, summarizes nearly 70 works at their intersections. We conc"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.05451","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/2506.05451/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":"2026-07-05T11:17:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"UebNb3nv/ydbAav6H/vuKua45J531J6XuDmkuuoyn5oje043IMbanNyrsv7CMpQcJOKqq72LgVRSZ2Um4bQqAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-19T03:27:47.910969Z"},"content_sha256":"b56f380205c6a12a44737e320ddb6a6eb983d528acccb1530c7c842bb59b0447","schema_version":"1.0","event_id":"sha256:b56f380205c6a12a44737e320ddb6a6eb983d528acccb1530c7c842bb59b0447"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/FQ5LFIOMJGGG54HJNL567P72SG/bundle.json","state_url":"https://pith.science/pith/FQ5LFIOMJGGG54HJNL567P72SG/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/FQ5LFIOMJGGG54HJNL567P72SG/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":"2026-08-19T03:27:47Z","links":{"resolver":"https://pith.science/pith/FQ5LFIOMJGGG54HJNL567P72SG","bundle":"https://pith.science/pith/FQ5LFIOMJGGG54HJNL567P72SG/bundle.json","state":"https://pith.science/pith/FQ5LFIOMJGGG54HJNL567P72SG/state.json","well_known_bundle":"https://pith.science/.well-known/pith/FQ5LFIOMJGGG54HJNL567P72SG/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:FQ5LFIOMJGGG54HJNL567P72SG","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":"3e991ade848579be124829cf6a58359c632238ed6fdcc86aebe639ee1bff998b","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2025-06-05T17:56:05Z","title_canon_sha256":"dd8e9f89d97cb1d95ca6225cc1ef07b22d51c7658e2c2506c155b9571da72c7e"},"schema_version":"1.0","source":{"id":"2506.05451","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.05451","created_at":"2026-07-05T11:17:02Z"},{"alias_kind":"arxiv_version","alias_value":"2506.05451v1","created_at":"2026-07-05T11:17:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.05451","created_at":"2026-07-05T11:17:02Z"},{"alias_kind":"pith_short_12","alias_value":"FQ5LFIOMJGGG","created_at":"2026-07-05T11:17:02Z"},{"alias_kind":"pith_short_16","alias_value":"FQ5LFIOMJGGG54HJ","created_at":"2026-07-05T11:17:02Z"},{"alias_kind":"pith_short_8","alias_value":"FQ5LFIOM","created_at":"2026-07-05T11:17:02Z"}],"graph_snapshots":[{"event_id":"sha256:b56f380205c6a12a44737e320ddb6a6eb983d528acccb1530c7c842bb59b0447","target":"graph","created_at":"2026-07-05T11:17:02Z","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/2506.05451/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"As large language models (LLMs) see wider real-world use, understanding and mitigating their unsafe behaviors is critical. Interpretation techniques can reveal causes of unsafe outputs and guide safety, but such connections with safety are often overlooked in prior surveys. We present the first survey that bridges this gap, introducing a unified framework that connects safety-focused interpretation methods, the safety enhancements they inform, and the tools that operationalize them. Our novel taxonomy, organized by LLM workflow stages, summarizes nearly 70 works at their intersections. We conc","authors_text":"Aeree Cho, Duen Horng Chau, Grace C. Kim, Mansi Phute, Seongmin Lee, ShengYun Peng","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2025-06-05T17:56:05Z","title":"Interpretation Meets Safety: A Survey on Interpretation Methods and Tools for Improving LLM Safety"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.05451","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:b05255df23ab4e5d15ae25e451e3475fd9e88066f5eb47aa9201aa5942af3ae6","target":"record","created_at":"2026-07-05T11:17:02Z","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":"3e991ade848579be124829cf6a58359c632238ed6fdcc86aebe639ee1bff998b","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.SE","submitted_at":"2025-06-05T17:56:05Z","title_canon_sha256":"dd8e9f89d97cb1d95ca6225cc1ef07b22d51c7658e2c2506c155b9571da72c7e"},"schema_version":"1.0","source":{"id":"2506.05451","kind":"arxiv","version":1}},"canonical_sha256":"2c3ab2a1cc498c6ef0e96afbefbffa91a8627b4c04bfab99bd9f69e8e261fdb4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2c3ab2a1cc498c6ef0e96afbefbffa91a8627b4c04bfab99bd9f69e8e261fdb4","first_computed_at":"2026-07-05T11:17:02.990526Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:17:02.990526Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"7QaAXFYo19rgIoc0JjchmIFjYcEaQFNsCKg5xBKYW6EGsCBr4MtNxKbFTYX5BmSmIFQ120QenH+vZgn5bhQBDw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:17:02.991005Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.05451","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b05255df23ab4e5d15ae25e451e3475fd9e88066f5eb47aa9201aa5942af3ae6","sha256:b56f380205c6a12a44737e320ddb6a6eb983d528acccb1530c7c842bb59b0447"],"state_sha256":"0de4a4b1a07f2e6103143dd8d717b8e1e45c5ae98c4b555ce4825a91fd03774c"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"htHg0Zhsf8xg+inAdLUMegwEQlsJu5csMBMep7x7333cOpzx8V+ESYUTY20qSplqETG/Qrhlb4Vguw5eS0V6BA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-19T03:27:47.916236Z","bundle_sha256":"d30a5ec639b73497421b89d2eda47efa80cf22c23320c3fc8a80c658a72caee0"}}