{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:7RIMH6J2WXYZUNEQHSVLLRSCTN","short_pith_number":"pith:7RIMH6J2","canonical_record":{"source":{"id":"2501.13126","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-21T13:12:13Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"f595cb3be1044f0a38a40283cf57b77769eb0b70532a5bf5effab47b71e8b032","abstract_canon_sha256":"30fc6fb7647cc467a8762190dfc71528e2007b775c064896449e6739f37e7435"},"schema_version":"1.0"},"canonical_sha256":"fc50c3f93ab5f19a34903caab5c6429b48fed2ecda1d9b07f7b1f95c168134d8","source":{"kind":"arxiv","id":"2501.13126","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.13126","created_at":"2026-07-05T10:15:24Z"},{"alias_kind":"arxiv_version","alias_value":"2501.13126v2","created_at":"2026-07-05T10:15:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.13126","created_at":"2026-07-05T10:15:24Z"},{"alias_kind":"pith_short_12","alias_value":"7RIMH6J2WXYZ","created_at":"2026-07-05T10:15:24Z"},{"alias_kind":"pith_short_16","alias_value":"7RIMH6J2WXYZUNEQ","created_at":"2026-07-05T10:15:24Z"},{"alias_kind":"pith_short_8","alias_value":"7RIMH6J2","created_at":"2026-07-05T10:15:24Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:7RIMH6J2WXYZUNEQHSVLLRSCTN","target":"record","payload":{"canonical_record":{"source":{"id":"2501.13126","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-21T13:12:13Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"f595cb3be1044f0a38a40283cf57b77769eb0b70532a5bf5effab47b71e8b032","abstract_canon_sha256":"30fc6fb7647cc467a8762190dfc71528e2007b775c064896449e6739f37e7435"},"schema_version":"1.0"},"canonical_sha256":"fc50c3f93ab5f19a34903caab5c6429b48fed2ecda1d9b07f7b1f95c168134d8","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:15:24.751625Z","signature_b64":"K0FoqeFppohWGAqfXlas1x/svZfAkhv1OyukW4u+h7xvG6hLEAHGuRzOdub8gG5nNEKRppAYfkOC/ZqnZM4QCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"fc50c3f93ab5f19a34903caab5c6429b48fed2ecda1d9b07f7b1f95c168134d8","last_reissued_at":"2026-07-05T10:15:24.751150Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:15:24.751150Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2501.13126","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":"2026-07-05T10:15:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NanZrPhxBZxO9rVtCH+KnMqfMEmJ0+xQDAvsgHgcT41Pcm+kWxDPKZyWh1d6stKuZqw3Qjs6WHgfot1huaEHAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T13:38:46.810167Z"},"content_sha256":"6c589c1c62dcde6e482cce2f511e9cfa083a87f7fa142ea7a0520cdd370067fb","schema_version":"1.0","event_id":"sha256:6c589c1c62dcde6e482cce2f511e9cfa083a87f7fa142ea7a0520cdd370067fb"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:7RIMH6J2WXYZUNEQHSVLLRSCTN","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Preference Curriculum: LLMs Should Always Be Pretrained on Their Preferred Data","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Feiyu Duan, Jingang Wang, Liangyu Xu, Rongxiang Weng, Sirui Wang, Xuemiao Zhang, Xunliang Cai, Yongwei Zhou","submitted_at":"2025-01-21T13:12:13Z","abstract_excerpt":"Large language models (LLMs) generally utilize a consistent data distribution throughout the pretraining process. However, as the model's capability improves, it is intuitive that its data preferences dynamically change, indicating the need for pretraining with different data at various training stages. To achieve it, we propose the Perplexity Difference (PD) based Preference Curriculum learning (PDPC) framework, which always perceives and uses the data preferred by LLMs to train and boost them. First, we introduce the PD metric to quantify the difference in how challenging a sample is for wea"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.13126","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/2501.13126/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":"2026-07-05T10:15:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/FFo812EDJALSQIWuhxoqVGp40lh9vc234cPtGYplAFbfYwLvAquVcH7KsFwLwzbQNlD8p1KUnMVuNlBGXMYDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T13:38:46.810554Z"},"content_sha256":"b438f8dfab8b8b6211a2e153f854b40cf7fac265b55b8ee8b191de49fec0c168","schema_version":"1.0","event_id":"sha256:b438f8dfab8b8b6211a2e153f854b40cf7fac265b55b8ee8b191de49fec0c168"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/7RIMH6J2WXYZUNEQHSVLLRSCTN/bundle.json","state_url":"https://pith.science/pith/7RIMH6J2WXYZUNEQHSVLLRSCTN/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/7RIMH6J2WXYZUNEQHSVLLRSCTN/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":"2026-07-07T13:38:46Z","links":{"resolver":"https://pith.science/pith/7RIMH6J2WXYZUNEQHSVLLRSCTN","bundle":"https://pith.science/pith/7RIMH6J2WXYZUNEQHSVLLRSCTN/bundle.json","state":"https://pith.science/pith/7RIMH6J2WXYZUNEQHSVLLRSCTN/state.json","well_known_bundle":"https://pith.science/.well-known/pith/7RIMH6J2WXYZUNEQHSVLLRSCTN/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:7RIMH6J2WXYZUNEQHSVLLRSCTN","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":"30fc6fb7647cc467a8762190dfc71528e2007b775c064896449e6739f37e7435","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-21T13:12:13Z","title_canon_sha256":"f595cb3be1044f0a38a40283cf57b77769eb0b70532a5bf5effab47b71e8b032"},"schema_version":"1.0","source":{"id":"2501.13126","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.13126","created_at":"2026-07-05T10:15:24Z"},{"alias_kind":"arxiv_version","alias_value":"2501.13126v2","created_at":"2026-07-05T10:15:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.13126","created_at":"2026-07-05T10:15:24Z"},{"alias_kind":"pith_short_12","alias_value":"7RIMH6J2WXYZ","created_at":"2026-07-05T10:15:24Z"},{"alias_kind":"pith_short_16","alias_value":"7RIMH6J2WXYZUNEQ","created_at":"2026-07-05T10:15:24Z"},{"alias_kind":"pith_short_8","alias_value":"7RIMH6J2","created_at":"2026-07-05T10:15:24Z"}],"graph_snapshots":[{"event_id":"sha256:b438f8dfab8b8b6211a2e153f854b40cf7fac265b55b8ee8b191de49fec0c168","target":"graph","created_at":"2026-07-05T10:15:24Z","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/2501.13126/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large language models (LLMs) generally utilize a consistent data distribution throughout the pretraining process. However, as the model's capability improves, it is intuitive that its data preferences dynamically change, indicating the need for pretraining with different data at various training stages. To achieve it, we propose the Perplexity Difference (PD) based Preference Curriculum learning (PDPC) framework, which always perceives and uses the data preferred by LLMs to train and boost them. First, we introduce the PD metric to quantify the difference in how challenging a sample is for wea","authors_text":"Feiyu Duan, Jingang Wang, Liangyu Xu, Rongxiang Weng, Sirui Wang, Xuemiao Zhang, Xunliang Cai, Yongwei Zhou","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-21T13:12:13Z","title":"Preference Curriculum: LLMs Should Always Be Pretrained on Their Preferred Data"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.13126","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":"sha256:6c589c1c62dcde6e482cce2f511e9cfa083a87f7fa142ea7a0520cdd370067fb","target":"record","created_at":"2026-07-05T10:15:24Z","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":"30fc6fb7647cc467a8762190dfc71528e2007b775c064896449e6739f37e7435","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-01-21T13:12:13Z","title_canon_sha256":"f595cb3be1044f0a38a40283cf57b77769eb0b70532a5bf5effab47b71e8b032"},"schema_version":"1.0","source":{"id":"2501.13126","kind":"arxiv","version":2}},"canonical_sha256":"fc50c3f93ab5f19a34903caab5c6429b48fed2ecda1d9b07f7b1f95c168134d8","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"fc50c3f93ab5f19a34903caab5c6429b48fed2ecda1d9b07f7b1f95c168134d8","first_computed_at":"2026-07-05T10:15:24.751150Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:15:24.751150Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"K0FoqeFppohWGAqfXlas1x/svZfAkhv1OyukW4u+h7xvG6hLEAHGuRzOdub8gG5nNEKRppAYfkOC/ZqnZM4QCg==","signature_status":"signed_v1","signed_at":"2026-07-05T10:15:24.751625Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.13126","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:6c589c1c62dcde6e482cce2f511e9cfa083a87f7fa142ea7a0520cdd370067fb","sha256:b438f8dfab8b8b6211a2e153f854b40cf7fac265b55b8ee8b191de49fec0c168"],"state_sha256":"4172b94315626d95c3e841988ffc5001f32c4695ae4139f144bc2eab024993ff"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"L32X8HIpkDkY0DwD4g2FBaWmDEDsOV6KzHF6gYjYm5JGuUOt/gXOAzK+C7hfq9OkR9YsAxI8S9K/dqX2a6DOBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T13:38:46.812475Z","bundle_sha256":"fc5ece3e1577542151dca874cc6b117ea0bac36101134dcdc5df459f184d9406"}}