{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:WNU5MOG5MQCF7D5MUUGDGMUSG2","short_pith_number":"pith:WNU5MOG5","schema_version":"1.0","canonical_sha256":"b369d638dd64045f8faca50c333292368e8f9d28913c7297bf5469beb13a08af","source":{"kind":"arxiv","id":"2507.11316","version":1},"attestation_state":"computed","paper":{"title":"Internal Value Alignment in Large Language Models through Controlled Value Vector Activation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Defu Lian, Haoran Jin, Meng Li, Minlie Huang, Xiting Wang, Yantao Jia, Zhihao Xu","submitted_at":"2025-07-15T13:48:35Z","abstract_excerpt":"Aligning Large Language Models (LLMs) with human values has attracted increasing attention since it provides clarity, transparency, and the ability to adapt to evolving scenarios. In this paper, we introduce a Controlled Value Vector Activation (ConVA) method that directly aligns the internal values of LLMs by interpreting how a value is encoded in their latent representations and modifies relevant activations to ensure consistent values in LLMs. To ensure an accurate and unbiased interpretation, we propose a context-controlled value vector identification method. To consistently control values"},"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":"2507.11316","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-07-15T13:48:35Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"e921be9302d344b3eb67735d73bacd62172ab6c0b35da5e73f8cb8d1faae49e0","abstract_canon_sha256":"264d2f06ca84f4712f6394f7e6d91ecf348bea11671dacedba7fcb5a40ad9389"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:37:36.895443Z","signature_b64":"cz+pD92O3bGKR/CMvsRKHnhUZjhTlswM740w4XwUbkkAgp3BY1bd/GOwMZ+3pI2pmh9xh2V2VWBwjeHDB2mJCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b369d638dd64045f8faca50c333292368e8f9d28913c7297bf5469beb13a08af","last_reissued_at":"2026-07-05T11:37:36.894967Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:37:36.894967Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Internal Value Alignment in Large Language Models through Controlled Value Vector Activation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Defu Lian, Haoran Jin, Meng Li, Minlie Huang, Xiting Wang, Yantao Jia, Zhihao Xu","submitted_at":"2025-07-15T13:48:35Z","abstract_excerpt":"Aligning Large Language Models (LLMs) with human values has attracted increasing attention since it provides clarity, transparency, and the ability to adapt to evolving scenarios. In this paper, we introduce a Controlled Value Vector Activation (ConVA) method that directly aligns the internal values of LLMs by interpreting how a value is encoded in their latent representations and modifies relevant activations to ensure consistent values in LLMs. To ensure an accurate and unbiased interpretation, we propose a context-controlled value vector identification method. To consistently control values"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2507.11316","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/2507.11316/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":"2507.11316","created_at":"2026-07-05T11:37:36.895031+00:00"},{"alias_kind":"arxiv_version","alias_value":"2507.11316v1","created_at":"2026-07-05T11:37:36.895031+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2507.11316","created_at":"2026-07-05T11:37:36.895031+00:00"},{"alias_kind":"pith_short_12","alias_value":"WNU5MOG5MQCF","created_at":"2026-07-05T11:37:36.895031+00:00"},{"alias_kind":"pith_short_16","alias_value":"WNU5MOG5MQCF7D5M","created_at":"2026-07-05T11:37:36.895031+00:00"},{"alias_kind":"pith_short_8","alias_value":"WNU5MOG5","created_at":"2026-07-05T11:37:36.895031+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/WNU5MOG5MQCF7D5MUUGDGMUSG2","json":"https://pith.science/pith/WNU5MOG5MQCF7D5MUUGDGMUSG2.json","graph_json":"https://pith.science/api/pith-number/WNU5MOG5MQCF7D5MUUGDGMUSG2/graph.json","events_json":"https://pith.science/api/pith-number/WNU5MOG5MQCF7D5MUUGDGMUSG2/events.json","paper":"https://pith.science/paper/WNU5MOG5"},"agent_actions":{"view_html":"https://pith.science/pith/WNU5MOG5MQCF7D5MUUGDGMUSG2","download_json":"https://pith.science/pith/WNU5MOG5MQCF7D5MUUGDGMUSG2.json","view_paper":"https://pith.science/paper/WNU5MOG5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2507.11316&json=true","fetch_graph":"https://pith.science/api/pith-number/WNU5MOG5MQCF7D5MUUGDGMUSG2/graph.json","fetch_events":"https://pith.science/api/pith-number/WNU5MOG5MQCF7D5MUUGDGMUSG2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WNU5MOG5MQCF7D5MUUGDGMUSG2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WNU5MOG5MQCF7D5MUUGDGMUSG2/action/storage_attestation","attest_author":"https://pith.science/pith/WNU5MOG5MQCF7D5MUUGDGMUSG2/action/author_attestation","sign_citation":"https://pith.science/pith/WNU5MOG5MQCF7D5MUUGDGMUSG2/action/citation_signature","submit_replication":"https://pith.science/pith/WNU5MOG5MQCF7D5MUUGDGMUSG2/action/replication_record"}},"created_at":"2026-07-05T11:37:36.895031+00:00","updated_at":"2026-07-05T11:37:36.895031+00:00"}