{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:BDHSE2VMRPD5DSPRCUTTCON5MW","short_pith_number":"pith:BDHSE2VM","canonical_record":{"source":{"id":"2406.18853","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-06-27T02:46:30Z","cross_cats_sorted":[],"title_canon_sha256":"e344a5566718daf508a545d2f1c79711222ba0b615841903c0b77174f8635051","abstract_canon_sha256":"3889915c68d51f42ba525327ba12a6bcf387d1c1a9d5a1dcd6566c9092d72dfb"},"schema_version":"1.0"},"canonical_sha256":"08cf226aac8bc7d1c9f115273139bd65b5dbfe5de06f251baed0bb4f70e4fd50","source":{"kind":"arxiv","id":"2406.18853","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.18853","created_at":"2026-07-05T09:27:02Z"},{"alias_kind":"arxiv_version","alias_value":"2406.18853v3","created_at":"2026-07-05T09:27:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.18853","created_at":"2026-07-05T09:27:02Z"},{"alias_kind":"pith_short_12","alias_value":"BDHSE2VMRPD5","created_at":"2026-07-05T09:27:02Z"},{"alias_kind":"pith_short_16","alias_value":"BDHSE2VMRPD5DSPR","created_at":"2026-07-05T09:27:02Z"},{"alias_kind":"pith_short_8","alias_value":"BDHSE2VM","created_at":"2026-07-05T09:27:02Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:BDHSE2VMRPD5DSPRCUTTCON5MW","target":"record","payload":{"canonical_record":{"source":{"id":"2406.18853","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-06-27T02:46:30Z","cross_cats_sorted":[],"title_canon_sha256":"e344a5566718daf508a545d2f1c79711222ba0b615841903c0b77174f8635051","abstract_canon_sha256":"3889915c68d51f42ba525327ba12a6bcf387d1c1a9d5a1dcd6566c9092d72dfb"},"schema_version":"1.0"},"canonical_sha256":"08cf226aac8bc7d1c9f115273139bd65b5dbfe5de06f251baed0bb4f70e4fd50","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:27:02.774303Z","signature_b64":"uAYqqwebeToLiHrd25R1L9vWHYCYo6O6OkJhSWojyNJtz58DQU9QQRu5REiRADPFPGOwEXvubl+HtfIv2kHsCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"08cf226aac8bc7d1c9f115273139bd65b5dbfe5de06f251baed0bb4f70e4fd50","last_reissued_at":"2026-07-05T09:27:02.773826Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:27:02.773826Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2406.18853","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:27:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xxUsaXmeQa2b73KEvRend5bx3zmHYvLIr8r1hczNfQZLrBm+qAM5FsTRuvVqKT7jpEI4uLTxaLvAv/U8KSP6AQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T08:20:56.702378Z"},"content_sha256":"b0ad4709a5a02c11e27a90c04f6c6a5c683d229c419407cb92d82ae10f72dd11","schema_version":"1.0","event_id":"sha256:b0ad4709a5a02c11e27a90c04f6c6a5c683d229c419407cb92d82ae10f72dd11"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:BDHSE2VMRPD5DSPRCUTTCON5MW","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Decoding-Time Language Model Alignment with Multiple Objectives","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Alisa Liu, Hannaneh Hajishirzi, Noah A. Smith, Ruizhe Shi, Simon S. Du, Yifang Chen, Yushi Hu","submitted_at":"2024-06-27T02:46:30Z","abstract_excerpt":"Aligning language models (LMs) to human preferences has emerged as a critical pursuit, enabling these models to better serve diverse user needs. Existing methods primarily focus on optimizing LMs for a single reward function, limiting their adaptability to varied objectives. Here, we propose $\\textbf{multi-objective decoding (MOD)}$, a decoding-time algorithm that outputs the next token from a linear combination of predictions of all base models, for any given weightings over different objectives. We exploit a common form among a family of $f$-divergence regularized alignment approaches (such "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.18853","kind":"arxiv","version":3},"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/2406.18853/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":"2026-07-05T09:27:02Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QwVL7Dighz02njeMdHlMcKFhKdzypyKN2T/r1/KGy8M+1wpdK4sLWUK8rQk60zgjLMtHYbIJoqVobFSLVN2hAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T08:20:56.702879Z"},"content_sha256":"1854d45179d292eb515a6a504c8b2ff511fad70c1688751d90f2318bafb2489a","schema_version":"1.0","event_id":"sha256:1854d45179d292eb515a6a504c8b2ff511fad70c1688751d90f2318bafb2489a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BDHSE2VMRPD5DSPRCUTTCON5MW/bundle.json","state_url":"https://pith.science/pith/BDHSE2VMRPD5DSPRCUTTCON5MW/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BDHSE2VMRPD5DSPRCUTTCON5MW/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":"2026-08-08T08:20:56Z","links":{"resolver":"https://pith.science/pith/BDHSE2VMRPD5DSPRCUTTCON5MW","bundle":"https://pith.science/pith/BDHSE2VMRPD5DSPRCUTTCON5MW/bundle.json","state":"https://pith.science/pith/BDHSE2VMRPD5DSPRCUTTCON5MW/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BDHSE2VMRPD5DSPRCUTTCON5MW/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:BDHSE2VMRPD5DSPRCUTTCON5MW","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":"3889915c68d51f42ba525327ba12a6bcf387d1c1a9d5a1dcd6566c9092d72dfb","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-06-27T02:46:30Z","title_canon_sha256":"e344a5566718daf508a545d2f1c79711222ba0b615841903c0b77174f8635051"},"schema_version":"1.0","source":{"id":"2406.18853","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.18853","created_at":"2026-07-05T09:27:02Z"},{"alias_kind":"arxiv_version","alias_value":"2406.18853v3","created_at":"2026-07-05T09:27:02Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.18853","created_at":"2026-07-05T09:27:02Z"},{"alias_kind":"pith_short_12","alias_value":"BDHSE2VMRPD5","created_at":"2026-07-05T09:27:02Z"},{"alias_kind":"pith_short_16","alias_value":"BDHSE2VMRPD5DSPR","created_at":"2026-07-05T09:27:02Z"},{"alias_kind":"pith_short_8","alias_value":"BDHSE2VM","created_at":"2026-07-05T09:27:02Z"}],"graph_snapshots":[{"event_id":"sha256:1854d45179d292eb515a6a504c8b2ff511fad70c1688751d90f2318bafb2489a","target":"graph","created_at":"2026-07-05T09:27:02Z","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/2406.18853/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Aligning language models (LMs) to human preferences has emerged as a critical pursuit, enabling these models to better serve diverse user needs. Existing methods primarily focus on optimizing LMs for a single reward function, limiting their adaptability to varied objectives. Here, we propose $\\textbf{multi-objective decoding (MOD)}$, a decoding-time algorithm that outputs the next token from a linear combination of predictions of all base models, for any given weightings over different objectives. We exploit a common form among a family of $f$-divergence regularized alignment approaches (such ","authors_text":"Alisa Liu, Hannaneh Hajishirzi, Noah A. Smith, Ruizhe Shi, Simon S. Du, Yifang Chen, Yushi Hu","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-06-27T02:46:30Z","title":"Decoding-Time Language Model Alignment with Multiple Objectives"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.18853","kind":"arxiv","version":3},"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:b0ad4709a5a02c11e27a90c04f6c6a5c683d229c419407cb92d82ae10f72dd11","target":"record","created_at":"2026-07-05T09:27:02Z","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":"3889915c68d51f42ba525327ba12a6bcf387d1c1a9d5a1dcd6566c9092d72dfb","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-06-27T02:46:30Z","title_canon_sha256":"e344a5566718daf508a545d2f1c79711222ba0b615841903c0b77174f8635051"},"schema_version":"1.0","source":{"id":"2406.18853","kind":"arxiv","version":3}},"canonical_sha256":"08cf226aac8bc7d1c9f115273139bd65b5dbfe5de06f251baed0bb4f70e4fd50","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"08cf226aac8bc7d1c9f115273139bd65b5dbfe5de06f251baed0bb4f70e4fd50","first_computed_at":"2026-07-05T09:27:02.773826Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:27:02.773826Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"uAYqqwebeToLiHrd25R1L9vWHYCYo6O6OkJhSWojyNJtz58DQU9QQRu5REiRADPFPGOwEXvubl+HtfIv2kHsCg==","signature_status":"signed_v1","signed_at":"2026-07-05T09:27:02.774303Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.18853","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b0ad4709a5a02c11e27a90c04f6c6a5c683d229c419407cb92d82ae10f72dd11","sha256:1854d45179d292eb515a6a504c8b2ff511fad70c1688751d90f2318bafb2489a"],"state_sha256":"d841a69c3528437a2e70e9fa4b11b8afe51fd8bd8a64dd9e0bed488a11981673"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NO65i69NtZtbgzfRMppKC0Q27JlEJiHziAWc4EqaKx4S+Ox5xT9TwRuWYsUXR6NYm1KH1ewzCOIr2jpI+dkOAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T08:20:56.708550Z","bundle_sha256":"764b1aa528d6ef1d30cc9ed0a4f1c9b2fcdd80792b98716e9573a9d8b00271b4"}}