{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:ZNXZZYVEO7LZTUA5ZIERP7OQ64","short_pith_number":"pith:ZNXZZYVE","canonical_record":{"source":{"id":"2402.02992","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-05T13:31:28Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"3fcc6e0d1e04192bdbd42fabb4936608abe6f70ae100c48a7f2176307bb53a59","abstract_canon_sha256":"c47c3e90a8ddec91830ca0c71ec1993230d2a0b6801876a21726faafe39b5605"},"schema_version":"1.0"},"canonical_sha256":"cb6f9ce2a477d799d01dca0917fdd0f72b9e6c5510a93882b4270ed775e13d6d","source":{"kind":"arxiv","id":"2402.02992","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.02992","created_at":"2026-07-05T08:22:36Z"},{"alias_kind":"arxiv_version","alias_value":"2402.02992v2","created_at":"2026-07-05T08:22:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.02992","created_at":"2026-07-05T08:22:36Z"},{"alias_kind":"pith_short_12","alias_value":"ZNXZZYVEO7LZ","created_at":"2026-07-05T08:22:36Z"},{"alias_kind":"pith_short_16","alias_value":"ZNXZZYVEO7LZTUA5","created_at":"2026-07-05T08:22:36Z"},{"alias_kind":"pith_short_8","alias_value":"ZNXZZYVE","created_at":"2026-07-05T08:22:36Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:ZNXZZYVEO7LZTUA5ZIERP7OQ64","target":"record","payload":{"canonical_record":{"source":{"id":"2402.02992","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-05T13:31:28Z","cross_cats_sorted":["cs.AI","cs.CL"],"title_canon_sha256":"3fcc6e0d1e04192bdbd42fabb4936608abe6f70ae100c48a7f2176307bb53a59","abstract_canon_sha256":"c47c3e90a8ddec91830ca0c71ec1993230d2a0b6801876a21726faafe39b5605"},"schema_version":"1.0"},"canonical_sha256":"cb6f9ce2a477d799d01dca0917fdd0f72b9e6c5510a93882b4270ed775e13d6d","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:22:36.681588Z","signature_b64":"lYlwtcX9BiJeK2a6C3MbW9iWkYT9F1Jf+1V+J43bERxyp7Z2SDyFAMmzau4TjvuiC+Sle+hk4e8PnYs4X3vJAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cb6f9ce2a477d799d01dca0917fdd0f72b9e6c5510a93882b4270ed775e13d6d","last_reissued_at":"2026-07-05T08:22:36.681040Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:22:36.681040Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2402.02992","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:22:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"UdWDJ1uouxkrQPkKlHRVOBv5cKKok5jtWLtc0anGwRipMOsK0GbAAgOJZvyA8Svc7jzNw549RD4gkqIMWZinBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T13:51:24.031768Z"},"content_sha256":"56d0a9e093b437b20dc1b6447de5d76b285282f08023b42a84927a52d8eb760b","schema_version":"1.0","event_id":"sha256:56d0a9e093b437b20dc1b6447de5d76b285282f08023b42a84927a52d8eb760b"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:ZNXZZYVEO7LZTUA5ZIERP7OQ64","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Decoding-time Realignment of Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL"],"primary_cat":"cs.LG","authors_text":"Daniele Calandriello, Felipe Llinares, Jessica Hoffmann, Leonardo Bianco, Lucas Dixon, Mathieu Blondel, Michal Valko, Quentin Berthet, Shangmin Guo, Tianlin Liu","submitted_at":"2024-02-05T13:31:28Z","abstract_excerpt":"Aligning language models with human preferences is crucial for reducing errors and biases in these models. Alignment techniques, such as reinforcement learning from human feedback (RLHF), are typically cast as optimizing a tradeoff between human preference rewards and a proximity regularization term that encourages staying close to the unaligned model. Selecting an appropriate level of regularization is critical: insufficient regularization can lead to reduced model capabilities due to reward hacking, whereas excessive regularization hinders alignment. Traditional methods for finding the optim"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.02992","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/2402.02992/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:22:36Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"rNnF34i0cHzitUqLQQ13YRoo4YqrrOnZysJtWCcJqVGWQmHHEkqAZKCHi1JA6VzrNf9zqEJCW0vEBwHI3hdrDQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-10T13:51:24.032393Z"},"content_sha256":"850c25a1480d36142bf195094443b43e54511e4fd03c2349733ad01aa9c170f8","schema_version":"1.0","event_id":"sha256:850c25a1480d36142bf195094443b43e54511e4fd03c2349733ad01aa9c170f8"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/ZNXZZYVEO7LZTUA5ZIERP7OQ64/bundle.json","state_url":"https://pith.science/pith/ZNXZZYVEO7LZTUA5ZIERP7OQ64/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/ZNXZZYVEO7LZTUA5ZIERP7OQ64/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-10T13:51:24Z","links":{"resolver":"https://pith.science/pith/ZNXZZYVEO7LZTUA5ZIERP7OQ64","bundle":"https://pith.science/pith/ZNXZZYVEO7LZTUA5ZIERP7OQ64/bundle.json","state":"https://pith.science/pith/ZNXZZYVEO7LZTUA5ZIERP7OQ64/state.json","well_known_bundle":"https://pith.science/.well-known/pith/ZNXZZYVEO7LZTUA5ZIERP7OQ64/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:ZNXZZYVEO7LZTUA5ZIERP7OQ64","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":"c47c3e90a8ddec91830ca0c71ec1993230d2a0b6801876a21726faafe39b5605","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-05T13:31:28Z","title_canon_sha256":"3fcc6e0d1e04192bdbd42fabb4936608abe6f70ae100c48a7f2176307bb53a59"},"schema_version":"1.0","source":{"id":"2402.02992","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2402.02992","created_at":"2026-07-05T08:22:36Z"},{"alias_kind":"arxiv_version","alias_value":"2402.02992v2","created_at":"2026-07-05T08:22:36Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.02992","created_at":"2026-07-05T08:22:36Z"},{"alias_kind":"pith_short_12","alias_value":"ZNXZZYVEO7LZ","created_at":"2026-07-05T08:22:36Z"},{"alias_kind":"pith_short_16","alias_value":"ZNXZZYVEO7LZTUA5","created_at":"2026-07-05T08:22:36Z"},{"alias_kind":"pith_short_8","alias_value":"ZNXZZYVE","created_at":"2026-07-05T08:22:36Z"}],"graph_snapshots":[{"event_id":"sha256:850c25a1480d36142bf195094443b43e54511e4fd03c2349733ad01aa9c170f8","target":"graph","created_at":"2026-07-05T08:22:36Z","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/2402.02992/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Aligning language models with human preferences is crucial for reducing errors and biases in these models. Alignment techniques, such as reinforcement learning from human feedback (RLHF), are typically cast as optimizing a tradeoff between human preference rewards and a proximity regularization term that encourages staying close to the unaligned model. Selecting an appropriate level of regularization is critical: insufficient regularization can lead to reduced model capabilities due to reward hacking, whereas excessive regularization hinders alignment. Traditional methods for finding the optim","authors_text":"Daniele Calandriello, Felipe Llinares, Jessica Hoffmann, Leonardo Bianco, Lucas Dixon, Mathieu Blondel, Michal Valko, Quentin Berthet, Shangmin Guo, Tianlin Liu","cross_cats":["cs.AI","cs.CL"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-05T13:31:28Z","title":"Decoding-time Realignment of Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.02992","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":"sha256:56d0a9e093b437b20dc1b6447de5d76b285282f08023b42a84927a52d8eb760b","target":"record","created_at":"2026-07-05T08:22:36Z","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":"c47c3e90a8ddec91830ca0c71ec1993230d2a0b6801876a21726faafe39b5605","cross_cats_sorted":["cs.AI","cs.CL"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-05T13:31:28Z","title_canon_sha256":"3fcc6e0d1e04192bdbd42fabb4936608abe6f70ae100c48a7f2176307bb53a59"},"schema_version":"1.0","source":{"id":"2402.02992","kind":"arxiv","version":2}},"canonical_sha256":"cb6f9ce2a477d799d01dca0917fdd0f72b9e6c5510a93882b4270ed775e13d6d","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"cb6f9ce2a477d799d01dca0917fdd0f72b9e6c5510a93882b4270ed775e13d6d","first_computed_at":"2026-07-05T08:22:36.681040Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:22:36.681040Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"lYlwtcX9BiJeK2a6C3MbW9iWkYT9F1Jf+1V+J43bERxyp7Z2SDyFAMmzau4TjvuiC+Sle+hk4e8PnYs4X3vJAA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:22:36.681588Z","signed_message":"canonical_sha256_bytes"},"source_id":"2402.02992","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:56d0a9e093b437b20dc1b6447de5d76b285282f08023b42a84927a52d8eb760b","sha256:850c25a1480d36142bf195094443b43e54511e4fd03c2349733ad01aa9c170f8"],"state_sha256":"c863f2854313e4d4b2317d23ca2a2bf1e683d4a93a26cddbf4c43598eb29aa2f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+32D4c+0Nceh2hHz+B+AssJU64/Q4mSUkPa9VhQLnfJDvFbt7voLOj2m9lrKFEuJYPnzrMAxK3gBv0in9EbGDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-10T13:51:24.037960Z","bundle_sha256":"43debeb19751260b65a0a4ad5cbb91987cad625e74fcce55d161881e65f879eb"}}