{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:DCTDCLZF2RWI2WYTLEWG5EV5MR","short_pith_number":"pith:DCTDCLZF","schema_version":"1.0","canonical_sha256":"18a6312f25d46c8d5b13592c6e92bd64561e9efc2c9051966823596423292aa1","source":{"kind":"arxiv","id":"2402.14433","version":1},"attestation_state":"computed","paper":{"title":"A Language Model's Guide Through Latent Space","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Dimitri von R\\\"utte, Gregor Bachmann, Sotiris Anagnostidis, Thomas Hofmann","submitted_at":"2024-02-22T10:25:14Z","abstract_excerpt":"Concept guidance has emerged as a cheap and simple way to control the behavior of language models by probing their hidden representations for concept vectors and using them to perturb activations at inference time. While the focus of previous work has largely been on truthfulness, in this paper we extend this framework to a richer set of concepts such as appropriateness, humor, creativity and quality, and explore to what degree current detection and guidance strategies work in these challenging settings. To facilitate evaluation, we develop a novel metric for concept guidance that takes into a"},"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":"2402.14433","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-22T10:25:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"760dd065d7c11f511a782b0c82048f3bf35c76571742ae83ea73c14a0d27a29d","abstract_canon_sha256":"de42a89b11f39bb7a0918ee73446f22a1e193db610a5705a71cfee0f4e4821f9"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:48:10.062589Z","signature_b64":"15LsaOt9MYxXbbYEZFNhQRAwaDc89X2AvUPkSXpj1nTuOn/y65wDpsfVBCeczvUH41PhdcRMY9MoUfUNEVLcBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"18a6312f25d46c8d5b13592c6e92bd64561e9efc2c9051966823596423292aa1","last_reissued_at":"2026-07-05T07:48:10.062086Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:48:10.062086Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Language Model's Guide Through Latent Space","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Dimitri von R\\\"utte, Gregor Bachmann, Sotiris Anagnostidis, Thomas Hofmann","submitted_at":"2024-02-22T10:25:14Z","abstract_excerpt":"Concept guidance has emerged as a cheap and simple way to control the behavior of language models by probing their hidden representations for concept vectors and using them to perturb activations at inference time. While the focus of previous work has largely been on truthfulness, in this paper we extend this framework to a richer set of concepts such as appropriateness, humor, creativity and quality, and explore to what degree current detection and guidance strategies work in these challenging settings. To facilitate evaluation, we develop a novel metric for concept guidance that takes into a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.14433","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/2402.14433/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":"2402.14433","created_at":"2026-07-05T07:48:10.062148+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.14433v1","created_at":"2026-07-05T07:48:10.062148+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.14433","created_at":"2026-07-05T07:48:10.062148+00:00"},{"alias_kind":"pith_short_12","alias_value":"DCTDCLZF2RWI","created_at":"2026-07-05T07:48:10.062148+00:00"},{"alias_kind":"pith_short_16","alias_value":"DCTDCLZF2RWI2WYT","created_at":"2026-07-05T07:48:10.062148+00:00"},{"alias_kind":"pith_short_8","alias_value":"DCTDCLZF","created_at":"2026-07-05T07:48:10.062148+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.30076","citing_title":"UniSteer: Text-Guided Flow Matching in Activation Space for Versatile LLM Steering","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2504.14668","citing_title":"A Byzantine Fault Tolerance Approach towards AI Safety","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2510.24941","citing_title":"Can Aha Moments Be Fake? Towards Quantifying Decorative and True Thinking in Chain-of-Thought","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2406.11717","citing_title":"Refusal in Language Models Is Mediated by a Single Direction","ref_index":193,"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":43,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DCTDCLZF2RWI2WYTLEWG5EV5MR","json":"https://pith.science/pith/DCTDCLZF2RWI2WYTLEWG5EV5MR.json","graph_json":"https://pith.science/api/pith-number/DCTDCLZF2RWI2WYTLEWG5EV5MR/graph.json","events_json":"https://pith.science/api/pith-number/DCTDCLZF2RWI2WYTLEWG5EV5MR/events.json","paper":"https://pith.science/paper/DCTDCLZF"},"agent_actions":{"view_html":"https://pith.science/pith/DCTDCLZF2RWI2WYTLEWG5EV5MR","download_json":"https://pith.science/pith/DCTDCLZF2RWI2WYTLEWG5EV5MR.json","view_paper":"https://pith.science/paper/DCTDCLZF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.14433&json=true","fetch_graph":"https://pith.science/api/pith-number/DCTDCLZF2RWI2WYTLEWG5EV5MR/graph.json","fetch_events":"https://pith.science/api/pith-number/DCTDCLZF2RWI2WYTLEWG5EV5MR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DCTDCLZF2RWI2WYTLEWG5EV5MR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DCTDCLZF2RWI2WYTLEWG5EV5MR/action/storage_attestation","attest_author":"https://pith.science/pith/DCTDCLZF2RWI2WYTLEWG5EV5MR/action/author_attestation","sign_citation":"https://pith.science/pith/DCTDCLZF2RWI2WYTLEWG5EV5MR/action/citation_signature","submit_replication":"https://pith.science/pith/DCTDCLZF2RWI2WYTLEWG5EV5MR/action/replication_record"}},"created_at":"2026-07-05T07:48:10.062148+00:00","updated_at":"2026-07-05T07:48:10.062148+00:00"}