{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:42BL6OBC6OVWOBRP7OVO5XPPLF","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":"f79008a3ba790a59ac19c9a897bc9e1dffeac194fec34474692e96afe961c0a3","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-06T05:47:04Z","title_canon_sha256":"44d73b59fe035a66f22095af86f20c1c7b98c9606bd4b923c264a29d2e026cab"},"schema_version":"1.0","source":{"id":"2507.04250","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2507.04250","created_at":"2026-07-05T11:32:44Z"},{"alias_kind":"arxiv_version","alias_value":"2507.04250v1","created_at":"2026-07-05T11:32:44Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.04250","created_at":"2026-07-05T11:32:44Z"},{"alias_kind":"pith_short_12","alias_value":"42BL6OBC6OVW","created_at":"2026-07-05T11:32:44Z"},{"alias_kind":"pith_short_16","alias_value":"42BL6OBC6OVWOBRP","created_at":"2026-07-05T11:32:44Z"},{"alias_kind":"pith_short_8","alias_value":"42BL6OBC","created_at":"2026-07-05T11:32:44Z"}],"graph_snapshots":[{"event_id":"sha256:89f4e0a1c30277cd50458e40e48dc4767b66f3d60f6b32388f0da9d6b9f61e5b","target":"graph","created_at":"2026-07-05T11:32:44Z","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/2507.04250/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Safety alignment is crucial for large language models (LLMs) to resist malicious instructions but often results in over-refusals, where benign prompts are unnecessarily rejected, impairing user experience and model utility. We introduce ACTOR (Activation-Based Training for Over-Refusal Reduction), a robust and compute- and data-efficient training framework that minimizes over-refusals by leveraging internal activation patterns from diverse queries. ACTOR precisely identifies and adjusts the activation components that trigger refusals, providing stronger control over the refusal mechanism. By f","authors_text":"Charles Fleming, Mahavir Dabas, Ming Jin, Ruoxi Jia, Si Chen","cross_cats":["cs.AI"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-06T05:47:04Z","title":"Just Enough Shifts: Mitigating Over-Refusal in Aligned Language Models with Targeted Representation Fine-Tuning"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.04250","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:fae0f9dcf1f9bf21c4622cb7ec88e32daed9a345662b750f971aeac718b63f9a","target":"record","created_at":"2026-07-05T11:32:44Z","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":"f79008a3ba790a59ac19c9a897bc9e1dffeac194fec34474692e96afe961c0a3","cross_cats_sorted":["cs.AI"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-07-06T05:47:04Z","title_canon_sha256":"44d73b59fe035a66f22095af86f20c1c7b98c9606bd4b923c264a29d2e026cab"},"schema_version":"1.0","source":{"id":"2507.04250","kind":"arxiv","version":1}},"canonical_sha256":"e682bf3822f3ab67062ffbaaeeddef597bde3c4c417ba2b3aaa70fca18e1890d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e682bf3822f3ab67062ffbaaeeddef597bde3c4c417ba2b3aaa70fca18e1890d","first_computed_at":"2026-07-05T11:32:44.226428Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:32:44.226428Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"HBIM4YiW3WGW6XZhEI6NJ3FoPsB532vT47ZEfWNM1OCbmf39zAvxXDatjVuEGUD3iXhmwSgwrm0vEKcRoB+eCw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:32:44.226900Z","signed_message":"canonical_sha256_bytes"},"source_id":"2507.04250","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:fae0f9dcf1f9bf21c4622cb7ec88e32daed9a345662b750f971aeac718b63f9a","sha256:89f4e0a1c30277cd50458e40e48dc4767b66f3d60f6b32388f0da9d6b9f61e5b"],"state_sha256":"c3c635cca387a7b0a019bd522d572e406d9fa2296bbdc8ebd6eea42b4bb7b49d"}