{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:DRVA5RXSKXWN5JYPI3WHPQB7TS","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"98d344ce776ae51bb95932688fc10a6149ff2030e711be48a925e2c03d1409ec","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-22T16:20:36Z","title_canon_sha256":"4fb93481aefc2f591965583e0278a3462cad892f21b3761154557eec74a7a579"},"schema_version":"1.0","source":{"id":"2404.14313","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2404.14313","created_at":"2026-07-05T08:21:15Z"},{"alias_kind":"arxiv_version","alias_value":"2404.14313v2","created_at":"2026-07-05T08:21:15Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.14313","created_at":"2026-07-05T08:21:15Z"},{"alias_kind":"pith_short_12","alias_value":"DRVA5RXSKXWN","created_at":"2026-07-05T08:21:15Z"},{"alias_kind":"pith_short_16","alias_value":"DRVA5RXSKXWN5JYP","created_at":"2026-07-05T08:21:15Z"},{"alias_kind":"pith_short_8","alias_value":"DRVA5RXS","created_at":"2026-07-05T08:21:15Z"}],"graph_snapshots":[{"event_id":"sha256:f08cdb0fbe968218ab5f824a4884c0d0b6f781d3a94a1338af2dff945e123b67","target":"graph","created_at":"2026-07-05T08:21:15Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2404.14313/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"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","authors_text":"Eric Zelikman, Jan-Philipp Fr\\\"anken, Kanishk Gandhi, Noah D. Goodman, Rafael Rafailov, Tobias Gerstenberg","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-22T16:20:36Z","title":"Self-Supervised Alignment with Mutual Information: Learning to Follow Principles without Preference Labels"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.14313","kind":"arxiv","version":2},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:851ff41558fc10e8eedb8b367140d98c67a9572502308df8e100796ba0b15017","target":"record","created_at":"2026-07-05T08:21:15Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"98d344ce776ae51bb95932688fc10a6149ff2030e711be48a925e2c03d1409ec","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-04-22T16:20:36Z","title_canon_sha256":"4fb93481aefc2f591965583e0278a3462cad892f21b3761154557eec74a7a579"},"schema_version":"1.0","source":{"id":"2404.14313","kind":"arxiv","version":2}},"canonical_sha256":"1c6a0ec6f255ecdea70f46ec77c03f9c8ba0f8afe12c9cdb7a0679fe2c119c42","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"1c6a0ec6f255ecdea70f46ec77c03f9c8ba0f8afe12c9cdb7a0679fe2c119c42","first_computed_at":"2026-07-05T08:21:15.427082Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:21:15.427082Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"nqeNzimT/7tdXXNR4DW1PKVXcGR5oNcSzFwR+cNtRmhnDV9J/hsgyrpwvSG5qbfeA+s1PKnRhmpIWyP/anv5BQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:21:15.427543Z","signed_message":"canonical_sha256_bytes"},"source_id":"2404.14313","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:851ff41558fc10e8eedb8b367140d98c67a9572502308df8e100796ba0b15017","sha256:f08cdb0fbe968218ab5f824a4884c0d0b6f781d3a94a1338af2dff945e123b67"],"state_sha256":"a22e9787ef5b501016d261c553d7a322fa947c47917109ce3e2986c33fc0da8a"}