{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:YY5I5FDZWY6NWJ7CUECVIZSTEK","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":"6bd4e96b270aab6cab31d377820ec3de8c1d90e4eb13dbfdf9cbcaf5fc20f5b8","cross_cats_sorted":["cs.SE"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-21T21:13:55Z","title_canon_sha256":"690bc91f9f98a8aee3d90fd3fd45ad83f4f7676aae0a3cf7d85c79cdbc7bb533"},"schema_version":"1.0","source":{"id":"2606.22673","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.22673","created_at":"2026-06-23T02:13:44Z"},{"alias_kind":"arxiv_version","alias_value":"2606.22673v1","created_at":"2026-06-23T02:13:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.22673","created_at":"2026-06-23T02:13:44Z"},{"alias_kind":"pith_short_12","alias_value":"YY5I5FDZWY6N","created_at":"2026-06-23T02:13:44Z"},{"alias_kind":"pith_short_16","alias_value":"YY5I5FDZWY6NWJ7C","created_at":"2026-06-23T02:13:44Z"},{"alias_kind":"pith_short_8","alias_value":"YY5I5FDZ","created_at":"2026-06-23T02:13:44Z"}],"graph_snapshots":[{"event_id":"sha256:bc58c35ffba518c1ea12200fc2ba264e82f4e4562ef98209cd8ed20efe0039be","target":"graph","created_at":"2026-06-23T02:13:44Z","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/2606.22673/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Coding agents based on large language models (LLMs) demonstrate remarkable autonomous capabilities, but they also introduce significant safety and misuse risks during multi-turn interactions with external environments. Existing safety mechanisms mainly rely on external guardrails, which have a limited ability to perform fine-grained behavioral control during execution. Meanwhile, recent mechanistic interpretability methods for LLM safety are mostly confined to single-turn or jailbreak-style QA settings, limiting their ability to capture the evolving risk dynamics of multi-turn agent execution.","authors_text":"Chaowei Xiao, Jie Cai, Jingcheng Niu, Mingyu Jin, Qiming Zhang, Weidi Luo, Yihao Quan, Zhen Xiang","cross_cats":["cs.SE"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-21T21:13:55Z","title":"AgentLens: Interpretable Safety Steering via Mechanistic Subspaces for Multi-Turn Coding Agent"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.22673","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":"sha256:48e9aa139cf95c6c020bd1b71b5dc2f8524485a0f369452e41bc70a1ef7db3b4","target":"record","created_at":"2026-06-23T02:13:44Z","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":"6bd4e96b270aab6cab31d377820ec3de8c1d90e4eb13dbfdf9cbcaf5fc20f5b8","cross_cats_sorted":["cs.SE"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-21T21:13:55Z","title_canon_sha256":"690bc91f9f98a8aee3d90fd3fd45ad83f4f7676aae0a3cf7d85c79cdbc7bb533"},"schema_version":"1.0","source":{"id":"2606.22673","kind":"arxiv","version":1}},"canonical_sha256":"c63a8e9479b63cdb27e2a105546653229fc67312cd1a93e9d0d7bdf95d5b12c1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c63a8e9479b63cdb27e2a105546653229fc67312cd1a93e9d0d7bdf95d5b12c1","first_computed_at":"2026-06-23T02:13:44.480201Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-23T02:13:44.480201Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"cEdUbTSo230zzOVXFlp4+uVTsf3y5H0K1ODGCuEsr00+eWNorr0x5MA31Js+MUXhgqQAkY0RMiHlYDh0DlRDBw==","signature_status":"signed_v1","signed_at":"2026-06-23T02:13:44.480627Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.22673","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:48e9aa139cf95c6c020bd1b71b5dc2f8524485a0f369452e41bc70a1ef7db3b4","sha256:bc58c35ffba518c1ea12200fc2ba264e82f4e4562ef98209cd8ed20efe0039be"],"state_sha256":"937c1a2e299e68aee758c745de490a9876b0336dc5581a8980934f2acb58c19f"}