{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:LTJG3RMFIUUULMSWU5UEKZKJHG","short_pith_number":"pith:LTJG3RMF","schema_version":"1.0","canonical_sha256":"5cd26dc585452945b256a76845654939a16c12df249f794dbd680d2ac32bb72e","source":{"kind":"arxiv","id":"2410.23054","version":2},"attestation_state":"computed","paper":{"title":"Controlling Language and Diffusion Models by Transporting Activations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Arno Blaas, Luca Zappella, Marco Cuturi, Michal Klein, Nicholas Apostoloff, Pau Rodriguez, Xavier Suau","submitted_at":"2024-10-30T14:21:33Z","abstract_excerpt":"The increasing capabilities of large generative models and their ever more widespread deployment have raised concerns about their reliability, safety, and potential misuse. To address these issues, recent works have proposed to control model generation by steering model activations in order to effectively induce or prevent the emergence of concepts or behaviors in the generated output. In this paper we introduce Activation Transport (AcT), a general framework to steer activations guided by optimal transport theory that generalizes many previous activation-steering works. AcT is modality-agnost"},"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":"2410.23054","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-30T14:21:33Z","cross_cats_sorted":["cs.AI","cs.CL","cs.CV"],"title_canon_sha256":"bc87dca042780c446bbb0fdba90f63bb9fa0c53729a972d55336a41c473b79af","abstract_canon_sha256":"239e759d0b57589eec1fdc0f41492b0ee8662e78d167894bf1990faa7bfe05e4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:39:09.738680Z","signature_b64":"BuIWv8ljuHTn+l3kDZAlJ0wnUTtSjIP+ZBBMT64dfWaIr1Jdc4EXqvoI7QmENmms8IrsJGfHe/pG6qd8569+DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5cd26dc585452945b256a76845654939a16c12df249f794dbd680d2ac32bb72e","last_reissued_at":"2026-07-05T09:39:09.738126Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:39:09.738126Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Controlling Language and Diffusion Models by Transporting Activations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.CV"],"primary_cat":"cs.LG","authors_text":"Arno Blaas, Luca Zappella, Marco Cuturi, Michal Klein, Nicholas Apostoloff, Pau Rodriguez, Xavier Suau","submitted_at":"2024-10-30T14:21:33Z","abstract_excerpt":"The increasing capabilities of large generative models and their ever more widespread deployment have raised concerns about their reliability, safety, and potential misuse. To address these issues, recent works have proposed to control model generation by steering model activations in order to effectively induce or prevent the emergence of concepts or behaviors in the generated output. In this paper we introduce Activation Transport (AcT), a general framework to steer activations guided by optimal transport theory that generalizes many previous activation-steering works. AcT is modality-agnost"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.23054","kind":"arxiv","version":2},"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/2410.23054/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":"2410.23054","created_at":"2026-07-05T09:39:09.738194+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.23054v2","created_at":"2026-07-05T09:39:09.738194+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.23054","created_at":"2026-07-05T09:39:09.738194+00:00"},{"alias_kind":"pith_short_12","alias_value":"LTJG3RMFIUUU","created_at":"2026-07-05T09:39:09.738194+00:00"},{"alias_kind":"pith_short_16","alias_value":"LTJG3RMFIUUULMSW","created_at":"2026-07-05T09:39:09.738194+00:00"},{"alias_kind":"pith_short_8","alias_value":"LTJG3RMF","created_at":"2026-07-05T09:39:09.738194+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.04775","citing_title":"Activation Steering of Video Generation Models via Reduced-Order Linear Optimal Control","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2502.06809","citing_title":"Neurons Speak in Ranges: Breaking Free from Discrete Neuronal Attribution","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09213","citing_title":"SHIFT: Steering Hidden Intermediates in Flow Transformers","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2604.14602","citing_title":"CausalDetox: Causal Head Selection and Intervention for Language Model Detoxification","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19018","citing_title":"Local Linearity of LLMs Enables Activation Steering via Model-Based Linear Optimal Control","ref_index":39,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LTJG3RMFIUUULMSWU5UEKZKJHG","json":"https://pith.science/pith/LTJG3RMFIUUULMSWU5UEKZKJHG.json","graph_json":"https://pith.science/api/pith-number/LTJG3RMFIUUULMSWU5UEKZKJHG/graph.json","events_json":"https://pith.science/api/pith-number/LTJG3RMFIUUULMSWU5UEKZKJHG/events.json","paper":"https://pith.science/paper/LTJG3RMF"},"agent_actions":{"view_html":"https://pith.science/pith/LTJG3RMFIUUULMSWU5UEKZKJHG","download_json":"https://pith.science/pith/LTJG3RMFIUUULMSWU5UEKZKJHG.json","view_paper":"https://pith.science/paper/LTJG3RMF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.23054&json=true","fetch_graph":"https://pith.science/api/pith-number/LTJG3RMFIUUULMSWU5UEKZKJHG/graph.json","fetch_events":"https://pith.science/api/pith-number/LTJG3RMFIUUULMSWU5UEKZKJHG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LTJG3RMFIUUULMSWU5UEKZKJHG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LTJG3RMFIUUULMSWU5UEKZKJHG/action/storage_attestation","attest_author":"https://pith.science/pith/LTJG3RMFIUUULMSWU5UEKZKJHG/action/author_attestation","sign_citation":"https://pith.science/pith/LTJG3RMFIUUULMSWU5UEKZKJHG/action/citation_signature","submit_replication":"https://pith.science/pith/LTJG3RMFIUUULMSWU5UEKZKJHG/action/replication_record"}},"created_at":"2026-07-05T09:39:09.738194+00:00","updated_at":"2026-07-05T09:39:09.738194+00:00"}