{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:LWFRW5DJQENXZPIB6PUESPOOB4","short_pith_number":"pith:LWFRW5DJ","canonical_record":{"source":{"id":"2410.01720","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-10-02T16:32:05Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"580a4c9c29885533914b67c66217e0661b45df2f6563ae428f19b561499645ce","abstract_canon_sha256":"a4e2db4f83e153c9ef9e527998dcd9aead143c74d27b4c6dc0b482cc8c2c3ba2"},"schema_version":"1.0"},"canonical_sha256":"5d8b1b7469811b7cbd01f3e8493dce0f0193a667fc79231368f0c18183edbd64","source":{"kind":"arxiv","id":"2410.01720","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.01720","created_at":"2026-07-05T10:10:09Z"},{"alias_kind":"arxiv_version","alias_value":"2410.01720v3","created_at":"2026-07-05T10:10:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.01720","created_at":"2026-07-05T10:10:09Z"},{"alias_kind":"pith_short_12","alias_value":"LWFRW5DJQENX","created_at":"2026-07-05T10:10:09Z"},{"alias_kind":"pith_short_16","alias_value":"LWFRW5DJQENXZPIB","created_at":"2026-07-05T10:10:09Z"},{"alias_kind":"pith_short_8","alias_value":"LWFRW5DJ","created_at":"2026-07-05T10:10:09Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:LWFRW5DJQENXZPIB6PUESPOOB4","target":"record","payload":{"canonical_record":{"source":{"id":"2410.01720","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-10-02T16:32:05Z","cross_cats_sorted":["cs.CL","cs.LG"],"title_canon_sha256":"580a4c9c29885533914b67c66217e0661b45df2f6563ae428f19b561499645ce","abstract_canon_sha256":"a4e2db4f83e153c9ef9e527998dcd9aead143c74d27b4c6dc0b482cc8c2c3ba2"},"schema_version":"1.0"},"canonical_sha256":"5d8b1b7469811b7cbd01f3e8493dce0f0193a667fc79231368f0c18183edbd64","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:10:09.175865Z","signature_b64":"maDfG974q2M5RIMz4j99rOBNwgL8hv86CbbuwFyuGoUJnd3tXcCgG9zZ1aHkc5JCAul2qYgSMaxVVeh3Yl7sBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5d8b1b7469811b7cbd01f3e8493dce0f0193a667fc79231368f0c18183edbd64","last_reissued_at":"2026-07-05T10:10:09.175446Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:10:09.175446Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.01720","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":"2026-07-05T10:10:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"8aUVtiN6N/FJyO3EH4QB1/TtY1JJ1XmXAvUjn+ZkWrvLLc8n9NjNY+49Y+VYq55LWQ5Q1Y4MiFZwsIbKYKR6BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T02:16:02.837742Z"},"content_sha256":"fb4162413b4237901c0248981b70dc0e990e7cce410a95e6f1a0d655ebf0d8bd","schema_version":"1.0","event_id":"sha256:fb4162413b4237901c0248981b70dc0e990e7cce410a95e6f1a0d655ebf0d8bd"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:LWFRW5DJQENXZPIB6PUESPOOB4","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Towards a Theoretical Understanding of Synthetic Data in LLM Post-Training: A Reverse-Bottleneck Perspective","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL","cs.LG"],"primary_cat":"cs.AI","authors_text":"Yong Liu, Zeyu Gan","submitted_at":"2024-10-02T16:32:05Z","abstract_excerpt":"Synthetic data has become a pivotal resource in post-training tasks for large language models (LLMs) due to the scarcity of high-quality, specific data. While various methods have been developed to generate synthetic data, there remains a discernible gap between the practical effects of synthetic data and our theoretical comprehension. To address this challenge, we commence by presenting a detailed modeling of the prevalent synthetic data generation process. Building upon this modeling, we demonstrate that the generalization capability of the post-trained model is critically determined by the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.01720","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/2410.01720/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":"2026-07-05T10:10:09Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"y9grgkh8fdX+fmuUdh/7WsfHfKtj7gXWrUr6Du7AlmC7hF/lGhyo3l70kL85i+mpABDY/EqGP1PWeK1bQDfwDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T02:16:02.838035Z"},"content_sha256":"d4c9577ea0a98b411d3e16b048bb9bc676184b9d35ed680c8c1bc9001975afb5","schema_version":"1.0","event_id":"sha256:d4c9577ea0a98b411d3e16b048bb9bc676184b9d35ed680c8c1bc9001975afb5"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/LWFRW5DJQENXZPIB6PUESPOOB4/bundle.json","state_url":"https://pith.science/pith/LWFRW5DJQENXZPIB6PUESPOOB4/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/LWFRW5DJQENXZPIB6PUESPOOB4/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":"2026-08-08T02:16:02Z","links":{"resolver":"https://pith.science/pith/LWFRW5DJQENXZPIB6PUESPOOB4","bundle":"https://pith.science/pith/LWFRW5DJQENXZPIB6PUESPOOB4/bundle.json","state":"https://pith.science/pith/LWFRW5DJQENXZPIB6PUESPOOB4/state.json","well_known_bundle":"https://pith.science/.well-known/pith/LWFRW5DJQENXZPIB6PUESPOOB4/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:LWFRW5DJQENXZPIB6PUESPOOB4","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":"a4e2db4f83e153c9ef9e527998dcd9aead143c74d27b4c6dc0b482cc8c2c3ba2","cross_cats_sorted":["cs.CL","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-10-02T16:32:05Z","title_canon_sha256":"580a4c9c29885533914b67c66217e0661b45df2f6563ae428f19b561499645ce"},"schema_version":"1.0","source":{"id":"2410.01720","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.01720","created_at":"2026-07-05T10:10:09Z"},{"alias_kind":"arxiv_version","alias_value":"2410.01720v3","created_at":"2026-07-05T10:10:09Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.01720","created_at":"2026-07-05T10:10:09Z"},{"alias_kind":"pith_short_12","alias_value":"LWFRW5DJQENX","created_at":"2026-07-05T10:10:09Z"},{"alias_kind":"pith_short_16","alias_value":"LWFRW5DJQENXZPIB","created_at":"2026-07-05T10:10:09Z"},{"alias_kind":"pith_short_8","alias_value":"LWFRW5DJ","created_at":"2026-07-05T10:10:09Z"}],"graph_snapshots":[{"event_id":"sha256:d4c9577ea0a98b411d3e16b048bb9bc676184b9d35ed680c8c1bc9001975afb5","target":"graph","created_at":"2026-07-05T10:10:09Z","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/2410.01720/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Synthetic data has become a pivotal resource in post-training tasks for large language models (LLMs) due to the scarcity of high-quality, specific data. While various methods have been developed to generate synthetic data, there remains a discernible gap between the practical effects of synthetic data and our theoretical comprehension. To address this challenge, we commence by presenting a detailed modeling of the prevalent synthetic data generation process. Building upon this modeling, we demonstrate that the generalization capability of the post-trained model is critically determined by the ","authors_text":"Yong Liu, Zeyu Gan","cross_cats":["cs.CL","cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-10-02T16:32:05Z","title":"Towards a Theoretical Understanding of Synthetic Data in LLM Post-Training: A Reverse-Bottleneck Perspective"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.01720","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:fb4162413b4237901c0248981b70dc0e990e7cce410a95e6f1a0d655ebf0d8bd","target":"record","created_at":"2026-07-05T10:10:09Z","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":"a4e2db4f83e153c9ef9e527998dcd9aead143c74d27b4c6dc0b482cc8c2c3ba2","cross_cats_sorted":["cs.CL","cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2024-10-02T16:32:05Z","title_canon_sha256":"580a4c9c29885533914b67c66217e0661b45df2f6563ae428f19b561499645ce"},"schema_version":"1.0","source":{"id":"2410.01720","kind":"arxiv","version":3}},"canonical_sha256":"5d8b1b7469811b7cbd01f3e8493dce0f0193a667fc79231368f0c18183edbd64","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"5d8b1b7469811b7cbd01f3e8493dce0f0193a667fc79231368f0c18183edbd64","first_computed_at":"2026-07-05T10:10:09.175446Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:10:09.175446Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"maDfG974q2M5RIMz4j99rOBNwgL8hv86CbbuwFyuGoUJnd3tXcCgG9zZ1aHkc5JCAul2qYgSMaxVVeh3Yl7sBw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:10:09.175865Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.01720","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:fb4162413b4237901c0248981b70dc0e990e7cce410a95e6f1a0d655ebf0d8bd","sha256:d4c9577ea0a98b411d3e16b048bb9bc676184b9d35ed680c8c1bc9001975afb5"],"state_sha256":"0da7ed2f9c6b1aa590c6b8b7f13da596288682e9211f7b243427fa6a26cbd3e4"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oIbfI9otwou2gl4dr31tlS+ryi4Ve7KtDob4Qeiteaf6KGS9Y2xN5v/DeN88LBU8SIr0Fs2aMKMvfmnFPNeRBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T02:16:02.841513Z","bundle_sha256":"c56fb2209f68d5508b845c04ec9fe36a84a753ad216a575187533475b247ab2a"}}