{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:L7UIQK7FASAFAPTV3OBFJICX4F","short_pith_number":"pith:L7UIQK7F","schema_version":"1.0","canonical_sha256":"5fe8882be50480503e75db8254a057e145761f671ddca3273a8e4ec2d0994a8c","source":{"kind":"arxiv","id":"2501.05764","version":1},"attestation_state":"computed","paper":{"title":"Controlling Large Language Models Through Concept Activation Vectors","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chengao Li, Hanyu Zhang, Qing He, Xiang Ao, Xiting Wang","submitted_at":"2025-01-10T07:41:48Z","abstract_excerpt":"As large language models (LLMs) are widely deployed across various domains, the ability to control their generated outputs has become more critical. This control involves aligning LLMs outputs with human values and ethical principles or customizing LLMs on specific topics or styles for individual users. Existing controlled generation methods either require significant computational resources and extensive trial-and-error or provide coarse-grained control. In this paper, we propose Generation with Concept Activation Vector (GCAV), a lightweight model control framework that ensures accurate cont"},"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":"2501.05764","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-10T07:41:48Z","cross_cats_sorted":[],"title_canon_sha256":"f1832da1bdc6f0a311490a0d630bb2e063fd78387a3a4769a85fcab44b98e67f","abstract_canon_sha256":"aca596fc9a5677ccaf4173274252c888d5c2c2451e9d9b0e9f02c49ecb5a72a3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:59:28.301411Z","signature_b64":"Ps2TFENACIXSA/mfnyqQLkTwIACcYTojYALzWITMkAnjcvIvwT6ipp4W1k1csmFAfNQiPur3wvyDKwdEOVk4DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5fe8882be50480503e75db8254a057e145761f671ddca3273a8e4ec2d0994a8c","last_reissued_at":"2026-07-05T09:59:28.300890Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:59:28.300890Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Controlling Large Language Models Through Concept Activation Vectors","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Chengao Li, Hanyu Zhang, Qing He, Xiang Ao, Xiting Wang","submitted_at":"2025-01-10T07:41:48Z","abstract_excerpt":"As large language models (LLMs) are widely deployed across various domains, the ability to control their generated outputs has become more critical. This control involves aligning LLMs outputs with human values and ethical principles or customizing LLMs on specific topics or styles for individual users. Existing controlled generation methods either require significant computational resources and extensive trial-and-error or provide coarse-grained control. In this paper, we propose Generation with Concept Activation Vector (GCAV), a lightweight model control framework that ensures accurate cont"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.05764","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/2501.05764/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":"2501.05764","created_at":"2026-07-05T09:59:28.300947+00:00"},{"alias_kind":"arxiv_version","alias_value":"2501.05764v1","created_at":"2026-07-05T09:59:28.300947+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.05764","created_at":"2026-07-05T09:59:28.300947+00:00"},{"alias_kind":"pith_short_12","alias_value":"L7UIQK7FASAF","created_at":"2026-07-05T09:59:28.300947+00:00"},{"alias_kind":"pith_short_16","alias_value":"L7UIQK7FASAFAPTV","created_at":"2026-07-05T09:59:28.300947+00:00"},{"alias_kind":"pith_short_8","alias_value":"L7UIQK7F","created_at":"2026-07-05T09:59:28.300947+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.12234","citing_title":"On The Effectiveness-Fluency Trade-Off In LLM Conditioning: A Systematic Study","ref_index":27,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/L7UIQK7FASAFAPTV3OBFJICX4F","json":"https://pith.science/pith/L7UIQK7FASAFAPTV3OBFJICX4F.json","graph_json":"https://pith.science/api/pith-number/L7UIQK7FASAFAPTV3OBFJICX4F/graph.json","events_json":"https://pith.science/api/pith-number/L7UIQK7FASAFAPTV3OBFJICX4F/events.json","paper":"https://pith.science/paper/L7UIQK7F"},"agent_actions":{"view_html":"https://pith.science/pith/L7UIQK7FASAFAPTV3OBFJICX4F","download_json":"https://pith.science/pith/L7UIQK7FASAFAPTV3OBFJICX4F.json","view_paper":"https://pith.science/paper/L7UIQK7F","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2501.05764&json=true","fetch_graph":"https://pith.science/api/pith-number/L7UIQK7FASAFAPTV3OBFJICX4F/graph.json","fetch_events":"https://pith.science/api/pith-number/L7UIQK7FASAFAPTV3OBFJICX4F/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/L7UIQK7FASAFAPTV3OBFJICX4F/action/timestamp_anchor","attest_storage":"https://pith.science/pith/L7UIQK7FASAFAPTV3OBFJICX4F/action/storage_attestation","attest_author":"https://pith.science/pith/L7UIQK7FASAFAPTV3OBFJICX4F/action/author_attestation","sign_citation":"https://pith.science/pith/L7UIQK7FASAFAPTV3OBFJICX4F/action/citation_signature","submit_replication":"https://pith.science/pith/L7UIQK7FASAFAPTV3OBFJICX4F/action/replication_record"}},"created_at":"2026-07-05T09:59:28.300947+00:00","updated_at":"2026-07-05T09:59:28.300947+00:00"}