{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:P7XNVAGUBCYSGQB2X276MP6WRB","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":"1b6d5c5c696c9bd94332b4c9cd2ab44a0a47a26c9418d9be51247585107227c5","cross_cats_sorted":["cs.LG","stat.ML"],"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-24T04:08:35Z","title_canon_sha256":"c0c7e29ba11458aaf3998f060dfdd0c732090b05c232d09a12f5e09e868c2f03"},"schema_version":"1.0","source":{"id":"2406.16306","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.16306","created_at":"2026-07-05T11:47:12Z"},{"alias_kind":"arxiv_version","alias_value":"2406.16306v3","created_at":"2026-07-05T11:47:12Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.16306","created_at":"2026-07-05T11:47:12Z"},{"alias_kind":"pith_short_12","alias_value":"P7XNVAGUBCYS","created_at":"2026-07-05T11:47:12Z"},{"alias_kind":"pith_short_16","alias_value":"P7XNVAGUBCYSGQB2","created_at":"2026-07-05T11:47:12Z"},{"alias_kind":"pith_short_8","alias_value":"P7XNVAGU","created_at":"2026-07-05T11:47:12Z"}],"graph_snapshots":[{"event_id":"sha256:ae18b406ab9dd8df2cf04a64f841e3bf11b8915270d68c9a049574b10a7e537c","target":"graph","created_at":"2026-07-05T11:47:12Z","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/2406.16306/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Aligning large language models (LLMs) with human preferences is essential for their applications. Recently, decoding-time alignment has emerged as an effective plug-and-play technique that avoids fine-tuning model parameters. This approach retains the general utility of pretrained LLMs but often suffers from significant inefficiencies during decoding, primarily due to wasted token generation and excessive reward evaluations. To address these challenges, we introduce Cascade Reward Sampling (CARDS) to resolve both efficiency bottlenecks in decoding-time alignment. Specifically, we develop a seg","authors_text":"Anamika Lochab, Ananth Grama, Bolian Li, Ruqi Zhang, Yifan Wang","cross_cats":["cs.LG","stat.ML"],"headline":"","license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-24T04:08:35Z","title":"Cascade Reward Sampling for Efficient Decoding-Time Alignment"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.16306","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:aec6ecd1c866dabba6bd6375de3f52ede49570834fcfa6d5359b76ea46d6b122","target":"record","created_at":"2026-07-05T11:47:12Z","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":"1b6d5c5c696c9bd94332b4c9cd2ab44a0a47a26c9418d9be51247585107227c5","cross_cats_sorted":["cs.LG","stat.ML"],"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CL","submitted_at":"2024-06-24T04:08:35Z","title_canon_sha256":"c0c7e29ba11458aaf3998f060dfdd0c732090b05c232d09a12f5e09e868c2f03"},"schema_version":"1.0","source":{"id":"2406.16306","kind":"arxiv","version":3}},"canonical_sha256":"7feeda80d408b123403abebfe63fd6885f3a2dfcf938d97d367e78deddc20924","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7feeda80d408b123403abebfe63fd6885f3a2dfcf938d97d367e78deddc20924","first_computed_at":"2026-07-05T11:47:12.379942Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:47:12.379942Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"CdMdMEoDRcdFBlbPZsGQYJ6mMzD9nzjyHMGL88RP5IWlBRh21Ksl5tyLw4eN6dHEkloVPsyNWIu5lqNuJ5SgDw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:47:12.380427Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.16306","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:aec6ecd1c866dabba6bd6375de3f52ede49570834fcfa6d5359b76ea46d6b122","sha256:ae18b406ab9dd8df2cf04a64f841e3bf11b8915270d68c9a049574b10a7e537c"],"state_sha256":"271b3990bbc907f8081be08b5ba578a6fa2a9b341a00291173ed6dddc15db377"}