{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:7Y5BE2D4R3I4FRNQ4H27BMQN74","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":"3084e01b39b6d57a577e4eebf94999b4facb4920a0e367e1a257be7dce6097b8","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-31T13:57:10Z","title_canon_sha256":"ba29ddc955920872fd7dcdafe1a291a83e37853b4201a687acff18a2bbbd2e56"},"schema_version":"1.0","source":{"id":"2506.00569","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2506.00569","created_at":"2026-07-05T11:13:37Z"},{"alias_kind":"arxiv_version","alias_value":"2506.00569v1","created_at":"2026-07-05T11:13:37Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.00569","created_at":"2026-07-05T11:13:37Z"},{"alias_kind":"pith_short_12","alias_value":"7Y5BE2D4R3I4","created_at":"2026-07-05T11:13:37Z"},{"alias_kind":"pith_short_16","alias_value":"7Y5BE2D4R3I4FRNQ","created_at":"2026-07-05T11:13:37Z"},{"alias_kind":"pith_short_8","alias_value":"7Y5BE2D4","created_at":"2026-07-05T11:13:37Z"}],"graph_snapshots":[{"event_id":"sha256:ed9de7cc570bcadf90fa3f5f13358eb6554f20f9d10b7e4c6a4c16f7cb2fd766","target":"graph","created_at":"2026-07-05T11:13:37Z","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/2506.00569/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"When aligning large language models (LLMs), their performance on various tasks (such as being helpful, harmless, and honest) depends heavily on the composition of their training data. However, selecting a data mixture that achieves strong performance across all tasks is challenging. Existing approaches rely on large ablation studies, heuristics, or human intuition, but these can be prohibitively expensive and suboptimal. We study this problem in the setting of preference optimization via DPO and introduce AutoMixAlign (AMA), a theoretically-grounded algorithm that adaptively mixes datasets dur","authors_text":"Adithya Devraj, Bing Yin, Chao Zhang, Hyokun Yun, Julian Katz-Samuels, Nicholas E. Corrado, Trishul Chilimbi, Yi Pan, Yi Xu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-31T13:57:10Z","title":"AutoMixAlign: Adaptive Data Mixing for Multi-Task Preference Optimization in LLMs"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.00569","kind":"arxiv","version":1},"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:c6305539b4fb193907ba62d331f67a856be2119e1c752e2d082019674145e966","target":"record","created_at":"2026-07-05T11:13:37Z","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":"3084e01b39b6d57a577e4eebf94999b4facb4920a0e367e1a257be7dce6097b8","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-05-31T13:57:10Z","title_canon_sha256":"ba29ddc955920872fd7dcdafe1a291a83e37853b4201a687acff18a2bbbd2e56"},"schema_version":"1.0","source":{"id":"2506.00569","kind":"arxiv","version":1}},"canonical_sha256":"fe3a12687c8ed1c2c5b0e1f5f0b20dff15c8e3c2abc8741b100626aaad36bc4c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"fe3a12687c8ed1c2c5b0e1f5f0b20dff15c8e3c2abc8741b100626aaad36bc4c","first_computed_at":"2026-07-05T11:13:37.060388Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:13:37.060388Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"eBJDH/jbBElaHD8VsXlHNj/gmqu+unaqWjT3sryPQ4tp6CTAhDcov4F0wpB/K1k8wOuBG7v+pSfeJ/Y0/3ZNAw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:13:37.060938Z","signed_message":"canonical_sha256_bytes"},"source_id":"2506.00569","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c6305539b4fb193907ba62d331f67a856be2119e1c752e2d082019674145e966","sha256:ed9de7cc570bcadf90fa3f5f13358eb6554f20f9d10b7e4c6a4c16f7cb2fd766"],"state_sha256":"42c24658c00e9cbec74ae3509e22e191afde9323fdb6e54ed84babf12f817817"}