{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UQHTWSVDNJBYTRBCNHQSQXO4ES","short_pith_number":"pith:UQHTWSVD","schema_version":"1.0","canonical_sha256":"a40f3b4aa36a4389c42269e1285ddc249c6874586eb5564a2d2905b855cc83d1","source":{"kind":"arxiv","id":"2606.05396","version":1},"attestation_state":"computed","paper":{"title":"Willing but Unable: Separating Refusal from Capability in Code LLMs via Abliteration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.SE"],"primary_cat":"cs.CR","authors_text":"Cristina Carleo, Domenico Cotroneo, Naghmeh Ivaki, Pietro Liguori","submitted_at":"2026-06-03T20:06:52Z","abstract_excerpt":"Producing a labeled vulnerable code at scale is a recurring obstacle for learning-based vulnerability detection: mined corpora carry substantial label noise, and existing LLM-based augmentation propagates these inaccuracies because it transforms vulnerable seeds rather than synthesising vulnerabilities from a specification. A complementary route is to start from safe code and ask an instruction-tuned LLM to inject a specified CWE (which would shift the labeling burden from open-ended detection to bounded binary confirmation) but safety-aligned code LLMs systematically refuse such prompts. This"},"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":"2606.05396","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CR","submitted_at":"2026-06-03T20:06:52Z","cross_cats_sorted":["cs.AI","cs.SE"],"title_canon_sha256":"0bf8d9ed447f57c90c1f2b71488959de531c09a2ebf457d4c1b5f51624ad4e3a","abstract_canon_sha256":"252cca11f3ce14acb60bf3d99b5af1498deb1b3baf8505d8e3ed6cf55d99d6f5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-05T00:13:56.298159Z","signature_b64":"5oJXn72smruOqZbubOhv9taSQ99QtMlUMWyBnAjYht0GNpLHvibG+euG9VSJDHcONZWHafeOS9ps2wbWbKKlBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a40f3b4aa36a4389c42269e1285ddc249c6874586eb5564a2d2905b855cc83d1","last_reissued_at":"2026-06-05T00:13:56.297751Z","signature_status":"signed_v1","first_computed_at":"2026-06-05T00:13:56.297751Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Willing but Unable: Separating Refusal from Capability in Code LLMs via Abliteration","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.SE"],"primary_cat":"cs.CR","authors_text":"Cristina Carleo, Domenico Cotroneo, Naghmeh Ivaki, Pietro Liguori","submitted_at":"2026-06-03T20:06:52Z","abstract_excerpt":"Producing a labeled vulnerable code at scale is a recurring obstacle for learning-based vulnerability detection: mined corpora carry substantial label noise, and existing LLM-based augmentation propagates these inaccuracies because it transforms vulnerable seeds rather than synthesising vulnerabilities from a specification. A complementary route is to start from safe code and ask an instruction-tuned LLM to inject a specified CWE (which would shift the labeling burden from open-ended detection to bounded binary confirmation) but safety-aligned code LLMs systematically refuse such prompts. This"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.05396","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/2606.05396/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":"2606.05396","created_at":"2026-06-05T00:13:56.297810+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.05396v1","created_at":"2026-06-05T00:13:56.297810+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.05396","created_at":"2026-06-05T00:13:56.297810+00:00"},{"alias_kind":"pith_short_12","alias_value":"UQHTWSVDNJBY","created_at":"2026-06-05T00:13:56.297810+00:00"},{"alias_kind":"pith_short_16","alias_value":"UQHTWSVDNJBYTRBC","created_at":"2026-06-05T00:13:56.297810+00:00"},{"alias_kind":"pith_short_8","alias_value":"UQHTWSVD","created_at":"2026-06-05T00:13:56.297810+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UQHTWSVDNJBYTRBCNHQSQXO4ES","json":"https://pith.science/pith/UQHTWSVDNJBYTRBCNHQSQXO4ES.json","graph_json":"https://pith.science/api/pith-number/UQHTWSVDNJBYTRBCNHQSQXO4ES/graph.json","events_json":"https://pith.science/api/pith-number/UQHTWSVDNJBYTRBCNHQSQXO4ES/events.json","paper":"https://pith.science/paper/UQHTWSVD"},"agent_actions":{"view_html":"https://pith.science/pith/UQHTWSVDNJBYTRBCNHQSQXO4ES","download_json":"https://pith.science/pith/UQHTWSVDNJBYTRBCNHQSQXO4ES.json","view_paper":"https://pith.science/paper/UQHTWSVD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.05396&json=true","fetch_graph":"https://pith.science/api/pith-number/UQHTWSVDNJBYTRBCNHQSQXO4ES/graph.json","fetch_events":"https://pith.science/api/pith-number/UQHTWSVDNJBYTRBCNHQSQXO4ES/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UQHTWSVDNJBYTRBCNHQSQXO4ES/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UQHTWSVDNJBYTRBCNHQSQXO4ES/action/storage_attestation","attest_author":"https://pith.science/pith/UQHTWSVDNJBYTRBCNHQSQXO4ES/action/author_attestation","sign_citation":"https://pith.science/pith/UQHTWSVDNJBYTRBCNHQSQXO4ES/action/citation_signature","submit_replication":"https://pith.science/pith/UQHTWSVDNJBYTRBCNHQSQXO4ES/action/replication_record"}},"created_at":"2026-06-05T00:13:56.297810+00:00","updated_at":"2026-06-05T00:13:56.297810+00:00"}