{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:37CAJHMWRZ32JLDHSVYZFHGXSK","short_pith_number":"pith:37CAJHMW","canonical_record":{"source":{"id":"2602.00747","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-01-31T14:27:46Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"cc20e42e088025237bff5873eb20923664f3e56c5c481ddcb6edd7cade0dc3a1","abstract_canon_sha256":"af9db8422dae007801afe8884835a25ac6540e61255c55198cd7df3d8e0d54b6"},"schema_version":"1.0"},"canonical_sha256":"dfc4049d968e77a4ac679571929cd792860787b43407b06d56f3195fb24d87db","source":{"kind":"arxiv","id":"2602.00747","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.00747","created_at":"2026-05-20T00:00:33Z"},{"alias_kind":"arxiv_version","alias_value":"2602.00747v2","created_at":"2026-05-20T00:00:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.00747","created_at":"2026-05-20T00:00:33Z"},{"alias_kind":"pith_short_12","alias_value":"37CAJHMWRZ32","created_at":"2026-05-20T00:00:33Z"},{"alias_kind":"pith_short_16","alias_value":"37CAJHMWRZ32JLDH","created_at":"2026-05-20T00:00:33Z"},{"alias_kind":"pith_short_8","alias_value":"37CAJHMW","created_at":"2026-05-20T00:00:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:37CAJHMWRZ32JLDHSVYZFHGXSK","target":"record","payload":{"canonical_record":{"source":{"id":"2602.00747","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-01-31T14:27:46Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"cc20e42e088025237bff5873eb20923664f3e56c5c481ddcb6edd7cade0dc3a1","abstract_canon_sha256":"af9db8422dae007801afe8884835a25ac6540e61255c55198cd7df3d8e0d54b6"},"schema_version":"1.0"},"canonical_sha256":"dfc4049d968e77a4ac679571929cd792860787b43407b06d56f3195fb24d87db","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:00:33.055927Z","signature_b64":"jtDlt+qQ8Zbsi9x3MtthQxy9xRtPk4QsPaBX90qJt2Ygr68d3shPF554z5CSWGklJucdaDgk2GImNtIfZHOHBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dfc4049d968e77a4ac679571929cd792860787b43407b06d56f3195fb24d87db","last_reissued_at":"2026-05-20T00:00:33.055263Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:00:33.055263Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2602.00747","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":"2026-05-20T00:00:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"25kQnj74qR3lfXmEIcnSE4bCKZRdiTht5c/GwD+RQDfFeJMsuG/bPxE5APeWRV/qgKZhHmZvy51/2i7SS3lXCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-27T10:20:14.364538Z"},"content_sha256":"3dc087c2129fe67b8efbd8358af7681ddcbd620a32a79d9553b5f406e362e8ab","schema_version":"1.0","event_id":"sha256:3dc087c2129fe67b8efbd8358af7681ddcbd620a32a79d9553b5f406e362e8ab"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:37CAJHMWRZ32JLDHSVYZFHGXSK","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Decouple Searching from Training: Scaling Data Mixing via Model Merging for Large Language Model Pre-training","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Fangcheng Shi, Fei Zhao, Haifeng Liu, Jieying Ye, Kaiyan Zhao, Shaosheng Cao, Shengrui Li, Yao Hu, Zheyong Xie","submitted_at":"2026-01-31T14:27:46Z","abstract_excerpt":"Determining an effective data mixture is a key factor in Large Language Model (LLM) pre-training, where models must balance general competence with proficiency on hard tasks such as math and code. However, identifying an optimal mixture remains an open challenge, as existing approaches either rely on unreliable tiny-scale proxy experiments or require prohibitively expensive large-scale exploration. To address this, we propose Decouple Searching from Training Mix (DeMix), a novel framework that leverages model merging to predict optimal data ratios. Instead of training proxy models for every sa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.00747","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/2602.00747/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":"2026-05-20T00:00:33Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"vDVfFA8ZF5CU5VGxcWYoZiOi0+sU1f+mILueQ8+FZG6ZqeR8TqxvN1JceqKT5O4WIQTmGIBjcTSIC8H54Z7gCw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-27T10:20:14.365024Z"},"content_sha256":"b0ba5738dafeec4f65202fb1471e5701a08b29d7a9db796be54d3d0305783c1a","schema_version":"1.0","event_id":"sha256:b0ba5738dafeec4f65202fb1471e5701a08b29d7a9db796be54d3d0305783c1a"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/37CAJHMWRZ32JLDHSVYZFHGXSK/bundle.json","state_url":"https://pith.science/pith/37CAJHMWRZ32JLDHSVYZFHGXSK/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/37CAJHMWRZ32JLDHSVYZFHGXSK/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":"2026-05-27T10:20:14Z","links":{"resolver":"https://pith.science/pith/37CAJHMWRZ32JLDHSVYZFHGXSK","bundle":"https://pith.science/pith/37CAJHMWRZ32JLDHSVYZFHGXSK/bundle.json","state":"https://pith.science/pith/37CAJHMWRZ32JLDHSVYZFHGXSK/state.json","well_known_bundle":"https://pith.science/.well-known/pith/37CAJHMWRZ32JLDHSVYZFHGXSK/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:37CAJHMWRZ32JLDHSVYZFHGXSK","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":"af9db8422dae007801afe8884835a25ac6540e61255c55198cd7df3d8e0d54b6","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-01-31T14:27:46Z","title_canon_sha256":"cc20e42e088025237bff5873eb20923664f3e56c5c481ddcb6edd7cade0dc3a1"},"schema_version":"1.0","source":{"id":"2602.00747","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2602.00747","created_at":"2026-05-20T00:00:33Z"},{"alias_kind":"arxiv_version","alias_value":"2602.00747v2","created_at":"2026-05-20T00:00:33Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2602.00747","created_at":"2026-05-20T00:00:33Z"},{"alias_kind":"pith_short_12","alias_value":"37CAJHMWRZ32","created_at":"2026-05-20T00:00:33Z"},{"alias_kind":"pith_short_16","alias_value":"37CAJHMWRZ32JLDH","created_at":"2026-05-20T00:00:33Z"},{"alias_kind":"pith_short_8","alias_value":"37CAJHMW","created_at":"2026-05-20T00:00:33Z"}],"graph_snapshots":[{"event_id":"sha256:b0ba5738dafeec4f65202fb1471e5701a08b29d7a9db796be54d3d0305783c1a","target":"graph","created_at":"2026-05-20T00:00:33Z","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/2602.00747/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Determining an effective data mixture is a key factor in Large Language Model (LLM) pre-training, where models must balance general competence with proficiency on hard tasks such as math and code. However, identifying an optimal mixture remains an open challenge, as existing approaches either rely on unreliable tiny-scale proxy experiments or require prohibitively expensive large-scale exploration. To address this, we propose Decouple Searching from Training Mix (DeMix), a novel framework that leverages model merging to predict optimal data ratios. Instead of training proxy models for every sa","authors_text":"Fangcheng Shi, Fei Zhao, Haifeng Liu, Jieying Ye, Kaiyan Zhao, Shaosheng Cao, Shengrui Li, Yao Hu, Zheyong Xie","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-01-31T14:27:46Z","title":"Decouple Searching from Training: Scaling Data Mixing via Model Merging for Large Language Model Pre-training"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2602.00747","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":"sha256:3dc087c2129fe67b8efbd8358af7681ddcbd620a32a79d9553b5f406e362e8ab","target":"record","created_at":"2026-05-20T00:00:33Z","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":"af9db8422dae007801afe8884835a25ac6540e61255c55198cd7df3d8e0d54b6","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2026-01-31T14:27:46Z","title_canon_sha256":"cc20e42e088025237bff5873eb20923664f3e56c5c481ddcb6edd7cade0dc3a1"},"schema_version":"1.0","source":{"id":"2602.00747","kind":"arxiv","version":2}},"canonical_sha256":"dfc4049d968e77a4ac679571929cd792860787b43407b06d56f3195fb24d87db","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"dfc4049d968e77a4ac679571929cd792860787b43407b06d56f3195fb24d87db","first_computed_at":"2026-05-20T00:00:33.055263Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-20T00:00:33.055263Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"jtDlt+qQ8Zbsi9x3MtthQxy9xRtPk4QsPaBX90qJt2Ygr68d3shPF554z5CSWGklJucdaDgk2GImNtIfZHOHBQ==","signature_status":"signed_v1","signed_at":"2026-05-20T00:00:33.055927Z","signed_message":"canonical_sha256_bytes"},"source_id":"2602.00747","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:3dc087c2129fe67b8efbd8358af7681ddcbd620a32a79d9553b5f406e362e8ab","sha256:b0ba5738dafeec4f65202fb1471e5701a08b29d7a9db796be54d3d0305783c1a"],"state_sha256":"00d6d65cbff51335766ecd1b0f91064fa17bf10b131512eced60b99f1a3047c6"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"w+/3Dx/3yA4Aqs6Ddn0XiIK7rbwQ0tbBFPFHAu80+7S24rT9tO7n/k2sllvgBayQCtLCB0qe1Sd0UPZPNqB+CA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-27T10:20:14.368535Z","bundle_sha256":"1f1b32d636ef38adc261572213cbac4ed37a43a1879e8f673af1b07fc7bb19e6"}}