{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:DRVA5RXSKXWN5JYPI3WHPQB7TS","short_pith_number":"pith:DRVA5RXS","schema_version":"1.0","canonical_sha256":"1c6a0ec6f255ecdea70f46ec77c03f9c8ba0f8afe12c9cdb7a0679fe2c119c42","source":{"kind":"arxiv","id":"2404.14313","version":2},"attestation_state":"computed","paper":{"title":"Self-Supervised Alignment with Mutual Information: Learning to Follow Principles without Preference Labels","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Eric Zelikman, Jan-Philipp Fr\\\"anken, Kanishk Gandhi, Noah D. Goodman, Rafael Rafailov, Tobias Gerstenberg","submitted_at":"2024-04-22T16:20:36Z","abstract_excerpt":"When prompting a language model (LM), users often expect the model to adhere to a set of behavioral principles across diverse tasks, such as producing insightful content while avoiding harmful or biased language. Instilling such principles (i.e., a constitution) into a model is resource-intensive, technically challenging, and generally requires human preference labels or examples. We introduce SAMI, an iterative algorithm that finetunes a pretrained language model (without requiring preference labels or demonstrations) to increase the conditional mutual information between constitutions and se"},"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":"2404.14313","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-22T16:20:36Z","cross_cats_sorted":[],"title_canon_sha256":"4fb93481aefc2f591965583e0278a3462cad892f21b3761154557eec74a7a579","abstract_canon_sha256":"98d344ce776ae51bb95932688fc10a6149ff2030e711be48a925e2c03d1409ec"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:21:15.427543Z","signature_b64":"nqeNzimT/7tdXXNR4DW1PKVXcGR5oNcSzFwR+cNtRmhnDV9J/hsgyrpwvSG5qbfeA+s1PKnRhmpIWyP/anv5BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"1c6a0ec6f255ecdea70f46ec77c03f9c8ba0f8afe12c9cdb7a0679fe2c119c42","last_reissued_at":"2026-07-05T08:21:15.427082Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:21:15.427082Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Self-Supervised Alignment with Mutual Information: Learning to Follow Principles without Preference Labels","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Eric Zelikman, Jan-Philipp Fr\\\"anken, Kanishk Gandhi, Noah D. Goodman, Rafael Rafailov, Tobias Gerstenberg","submitted_at":"2024-04-22T16:20:36Z","abstract_excerpt":"When prompting a language model (LM), users often expect the model to adhere to a set of behavioral principles across diverse tasks, such as producing insightful content while avoiding harmful or biased language. Instilling such principles (i.e., a constitution) into a model is resource-intensive, technically challenging, and generally requires human preference labels or examples. We introduce SAMI, an iterative algorithm that finetunes a pretrained language model (without requiring preference labels or demonstrations) to increase the conditional mutual information between constitutions and se"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.14313","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/2404.14313/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":"2404.14313","created_at":"2026-07-05T08:21:15.427138+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.14313v2","created_at":"2026-07-05T08:21:15.427138+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.14313","created_at":"2026-07-05T08:21:15.427138+00:00"},{"alias_kind":"pith_short_12","alias_value":"DRVA5RXSKXWN","created_at":"2026-07-05T08:21:15.427138+00:00"},{"alias_kind":"pith_short_16","alias_value":"DRVA5RXSKXWN5JYP","created_at":"2026-07-05T08:21:15.427138+00:00"},{"alias_kind":"pith_short_8","alias_value":"DRVA5RXS","created_at":"2026-07-05T08:21:15.427138+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.12923","citing_title":"Order Is Not Control","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DRVA5RXSKXWN5JYPI3WHPQB7TS","json":"https://pith.science/pith/DRVA5RXSKXWN5JYPI3WHPQB7TS.json","graph_json":"https://pith.science/api/pith-number/DRVA5RXSKXWN5JYPI3WHPQB7TS/graph.json","events_json":"https://pith.science/api/pith-number/DRVA5RXSKXWN5JYPI3WHPQB7TS/events.json","paper":"https://pith.science/paper/DRVA5RXS"},"agent_actions":{"view_html":"https://pith.science/pith/DRVA5RXSKXWN5JYPI3WHPQB7TS","download_json":"https://pith.science/pith/DRVA5RXSKXWN5JYPI3WHPQB7TS.json","view_paper":"https://pith.science/paper/DRVA5RXS","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.14313&json=true","fetch_graph":"https://pith.science/api/pith-number/DRVA5RXSKXWN5JYPI3WHPQB7TS/graph.json","fetch_events":"https://pith.science/api/pith-number/DRVA5RXSKXWN5JYPI3WHPQB7TS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DRVA5RXSKXWN5JYPI3WHPQB7TS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DRVA5RXSKXWN5JYPI3WHPQB7TS/action/storage_attestation","attest_author":"https://pith.science/pith/DRVA5RXSKXWN5JYPI3WHPQB7TS/action/author_attestation","sign_citation":"https://pith.science/pith/DRVA5RXSKXWN5JYPI3WHPQB7TS/action/citation_signature","submit_replication":"https://pith.science/pith/DRVA5RXSKXWN5JYPI3WHPQB7TS/action/replication_record"}},"created_at":"2026-07-05T08:21:15.427138+00:00","updated_at":"2026-07-05T08:21:15.427138+00:00"}