{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:TAWQTPNYV3AMHTCHDSVN6WNT37","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":"79e00a97f82f0b38e2c27b11d25c228c033cf5b659de1245806df1d6efca7566","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-07-23T07:56:05Z","title_canon_sha256":"2460229338a306cbb6e86695ea4031373304a42bbce55f597dee9a3a0ac6c727"},"schema_version":"1.0","source":{"id":"2507.21141","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.21141","created_at":"2026-07-05T11:44:32Z"},{"alias_kind":"arxiv_version","alias_value":"2507.21141v1","created_at":"2026-07-05T11:44:32Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.21141","created_at":"2026-07-05T11:44:32Z"},{"alias_kind":"pith_short_12","alias_value":"TAWQTPNYV3AM","created_at":"2026-07-05T11:44:32Z"},{"alias_kind":"pith_short_16","alias_value":"TAWQTPNYV3AMHTCH","created_at":"2026-07-05T11:44:32Z"},{"alias_kind":"pith_short_8","alias_value":"TAWQTPNY","created_at":"2026-07-05T11:44:32Z"}],"graph_snapshots":[{"event_id":"sha256:cc11aafcb86170adccee41fa61c8e3276ef10bb2766d0ba8e0e6778b9060f9d8","target":"graph","created_at":"2026-07-05T11:44:32Z","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/2507.21141/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Recent advances in large language models (LLMs) have intensified the need to understand and reliably curb their harmful behaviours. We introduce a multidimensional framework for probing and steering harmful content in model internals. For each of 55 distinct harmfulness subconcepts (e.g., racial hate, employment scams, weapons), we learn a linear probe, yielding 55 interpretable directions in activation space. Collectively, these directions span a harmfulness subspace that we show is strikingly low-rank. We then test ablation of the entire subspace from model internals, as well as steering and","authors_text":"Adhitya Rajendra Kumar, Kevin Zhu, McNair Shah, Naitik Chheda, Saleena Angeline, Sean O'Brien, Vasu Sharma, Will Cai","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-07-23T07:56:05Z","title":"The Geometry of Harmfulness in LLMs through Subconcept Probing"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.21141","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:be8e1eb5aa538c7b8963291b2546dd829c318e819b3f7d6aa474edd152d85af6","target":"record","created_at":"2026-07-05T11:44:32Z","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":"79e00a97f82f0b38e2c27b11d25c228c033cf5b659de1245806df1d6efca7566","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-07-23T07:56:05Z","title_canon_sha256":"2460229338a306cbb6e86695ea4031373304a42bbce55f597dee9a3a0ac6c727"},"schema_version":"1.0","source":{"id":"2507.21141","kind":"arxiv","version":1}},"canonical_sha256":"982d09bdb8aec0c3cc471caadf59b3dfef91a9052a4c4837ee5a4bb5d5af54ea","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"982d09bdb8aec0c3cc471caadf59b3dfef91a9052a4c4837ee5a4bb5d5af54ea","first_computed_at":"2026-07-05T11:44:32.094454Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:44:32.094454Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"iJhDjNFDq8cp01u5t6tM3HeDBpykP8Tkhtg1Z7psUu/I9EG1G7kwsODDIizyK1jiMqQW724Y35NZLWU721gtCw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:44:32.094951Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.21141","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:be8e1eb5aa538c7b8963291b2546dd829c318e819b3f7d6aa474edd152d85af6","sha256:cc11aafcb86170adccee41fa61c8e3276ef10bb2766d0ba8e0e6778b9060f9d8"],"state_sha256":"d4d5d0f4f510f0bc842c64a34a50dcbb4d5a0d64c3c784a06556580af9ff8241"}