{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:52XRS253GZNCZW4YPS37E7YNME","short_pith_number":"pith:52XRS253","schema_version":"1.0","canonical_sha256":"eeaf196bbb365a2cdb987cb7f27f0d610fab61f39daa5e93fd6e68f8bc21e5fb","source":{"kind":"arxiv","id":"2505.03189","version":1},"attestation_state":"computed","paper":{"title":"Patterns and Mechanisms of Contrastive Activation Engineering","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.HC"],"primary_cat":"cs.AI","authors_text":"Ayush Panda, Sheikh Abdur Raheem Ali, Stepan Shabalin, Yixiong Hao","submitted_at":"2025-05-06T05:15:12Z","abstract_excerpt":"Controlling the behavior of Large Language Models (LLMs) remains a significant challenge due to their inherent complexity and opacity. While techniques like fine-tuning can modify model behavior, they typically require extensive computational resources. Recent work has introduced a class of contrastive activation engineering (CAE) techniques as promising approaches for steering LLM outputs through targeted modifications to their internal representations. Applied at inference-time with zero cost, CAE has the potential to introduce a new paradigm of flexible, task-specific LLM behavior tuning. W"},"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":"2505.03189","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-05-06T05:15:12Z","cross_cats_sorted":["cs.HC"],"title_canon_sha256":"7186c395896999309abc3696f87c66cb43a9a37d17bea6266d34402e409a3e25","abstract_canon_sha256":"1cbe0abf4e9c35e21fd063f69c2f17f927e3b6a4ddd6e4e3a70ac247e4209212"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:59:11.792801Z","signature_b64":"UTrSMZVaaQGWNntSqN2ufbM2x9qf5knNkjsrnThD28weSJnvPjTBUz3jrOcEf801OqenVJJUNOFiPYmB63SsAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"eeaf196bbb365a2cdb987cb7f27f0d610fab61f39daa5e93fd6e68f8bc21e5fb","last_reissued_at":"2026-07-05T10:59:11.792324Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:59:11.792324Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Patterns and Mechanisms of Contrastive Activation Engineering","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.HC"],"primary_cat":"cs.AI","authors_text":"Ayush Panda, Sheikh Abdur Raheem Ali, Stepan Shabalin, Yixiong Hao","submitted_at":"2025-05-06T05:15:12Z","abstract_excerpt":"Controlling the behavior of Large Language Models (LLMs) remains a significant challenge due to their inherent complexity and opacity. While techniques like fine-tuning can modify model behavior, they typically require extensive computational resources. Recent work has introduced a class of contrastive activation engineering (CAE) techniques as promising approaches for steering LLM outputs through targeted modifications to their internal representations. Applied at inference-time with zero cost, CAE has the potential to introduce a new paradigm of flexible, task-specific LLM behavior tuning. W"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.03189","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/2505.03189/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":"2505.03189","created_at":"2026-07-05T10:59:11.792383+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.03189v1","created_at":"2026-07-05T10:59:11.792383+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.03189","created_at":"2026-07-05T10:59:11.792383+00:00"},{"alias_kind":"pith_short_12","alias_value":"52XRS253GZNC","created_at":"2026-07-05T10:59:11.792383+00:00"},{"alias_kind":"pith_short_16","alias_value":"52XRS253GZNCZW4Y","created_at":"2026-07-05T10:59:11.792383+00:00"},{"alias_kind":"pith_short_8","alias_value":"52XRS253","created_at":"2026-07-05T10:59:11.792383+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.03121","citing_title":"An Independent Safety Evaluation of Kimi K2.5","ref_index":97,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/52XRS253GZNCZW4YPS37E7YNME","json":"https://pith.science/pith/52XRS253GZNCZW4YPS37E7YNME.json","graph_json":"https://pith.science/api/pith-number/52XRS253GZNCZW4YPS37E7YNME/graph.json","events_json":"https://pith.science/api/pith-number/52XRS253GZNCZW4YPS37E7YNME/events.json","paper":"https://pith.science/paper/52XRS253"},"agent_actions":{"view_html":"https://pith.science/pith/52XRS253GZNCZW4YPS37E7YNME","download_json":"https://pith.science/pith/52XRS253GZNCZW4YPS37E7YNME.json","view_paper":"https://pith.science/paper/52XRS253","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.03189&json=true","fetch_graph":"https://pith.science/api/pith-number/52XRS253GZNCZW4YPS37E7YNME/graph.json","fetch_events":"https://pith.science/api/pith-number/52XRS253GZNCZW4YPS37E7YNME/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/52XRS253GZNCZW4YPS37E7YNME/action/timestamp_anchor","attest_storage":"https://pith.science/pith/52XRS253GZNCZW4YPS37E7YNME/action/storage_attestation","attest_author":"https://pith.science/pith/52XRS253GZNCZW4YPS37E7YNME/action/author_attestation","sign_citation":"https://pith.science/pith/52XRS253GZNCZW4YPS37E7YNME/action/citation_signature","submit_replication":"https://pith.science/pith/52XRS253GZNCZW4YPS37E7YNME/action/replication_record"}},"created_at":"2026-07-05T10:59:11.792383+00:00","updated_at":"2026-07-05T10:59:11.792383+00:00"}