{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2022:42FNTVNGRGPDPGCLDZRLBGDJOZ","short_pith_number":"pith:42FNTVNG","canonical_record":{"source":{"id":"2207.14255","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-07-28T17:40:47Z","cross_cats_sorted":[],"title_canon_sha256":"e3f17dec2398ee2bed0424c1881217774b214a1303ffa79fbf895246d41b0e20","abstract_canon_sha256":"5c730c8b906cbb6fa41ef63a66f39d68103c6ce78288ef7e6e78d63b08517e89"},"schema_version":"1.0"},"canonical_sha256":"e68ad9d5a6899e37984b1e62b098697672f8bfa48de329b4fb0bf6ca0b714960","source":{"kind":"arxiv","id":"2207.14255","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2207.14255","created_at":"2026-05-18T00:35:37Z"},{"alias_kind":"arxiv_version","alias_value":"2207.14255v1","created_at":"2026-05-18T00:35:37Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2207.14255","created_at":"2026-05-18T00:35:37Z"},{"alias_kind":"pith_short_12","alias_value":"42FNTVNGRGPD","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"42FNTVNGRGPDPGCL","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"42FNTVNG","created_at":"2026-05-18T12:33:33Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2022:42FNTVNGRGPDPGCLDZRLBGDJOZ","target":"record","payload":{"canonical_record":{"source":{"id":"2207.14255","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-07-28T17:40:47Z","cross_cats_sorted":[],"title_canon_sha256":"e3f17dec2398ee2bed0424c1881217774b214a1303ffa79fbf895246d41b0e20","abstract_canon_sha256":"5c730c8b906cbb6fa41ef63a66f39d68103c6ce78288ef7e6e78d63b08517e89"},"schema_version":"1.0"},"canonical_sha256":"e68ad9d5a6899e37984b1e62b098697672f8bfa48de329b4fb0bf6ca0b714960","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:35:37.704631Z","signature_b64":"UooRznMuMzlmbRmNk3htWg0hBm1b4o5py8Gb+xffuP/IkwGxj9Gh+/U1Px0jD/L6qf6SjCvIhg78XKpVzIEuCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e68ad9d5a6899e37984b1e62b098697672f8bfa48de329b4fb0bf6ca0b714960","last_reissued_at":"2026-05-18T00:35:37.704002Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:35:37.704002Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2207.14255","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T00:35:37Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nch8wwRnUtbkc5FYTOiXUoJFj/69CyGM3Q9o+3PGsudGJ33rDASmgEsXL4TIH7K3/YAul/cUwvFGvmFEnrBdBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-10T12:00:57.515495Z"},"content_sha256":"f79297fe33015ae3e4987d69b3f797f061d8d312ff2572b1fc777c5c875f6f9c","schema_version":"1.0","event_id":"sha256:f79297fe33015ae3e4987d69b3f797f061d8d312ff2572b1fc777c5c875f6f9c"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2022:42FNTVNGRGPDPGCLDZRLBGDJOZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Efficient Training of Language Models to Fill in the Middle","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Christine McLeavey, Heewoo Jun, Jerry Tworek, John Schulman, Mark Chen, Mohammad Bavarian, Nikolas Tezak","submitted_at":"2022-07-28T17:40:47Z","abstract_excerpt":"We show that autoregressive language models can learn to infill text after we apply a straightforward transformation to the dataset, which simply moves a span of text from the middle of a document to its end. While this data augmentation has garnered much interest in recent years, we provide extensive evidence that training models with a large fraction of data transformed in this way does not harm the original left-to-right generative capability, as measured by perplexity and sampling evaluations across a wide range of scales. Given the usefulness, simplicity, and efficiency of training models"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2207.14255","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"2026-05-18T00:35:37Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nBujjk5q3FjYGA2y7oegY0elGYAUcadOPKE5wGjph1UDgwF/0vKTeVX6CoJtL1gAGxKMruLti96zAnFSbWKYDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-10T12:00:57.516204Z"},"content_sha256":"352c7db04920b43f2d7e565520e264cd766c565e56badc4c4d89f8b2a9a5868d","schema_version":"1.0","event_id":"sha256:352c7db04920b43f2d7e565520e264cd766c565e56badc4c4d89f8b2a9a5868d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/42FNTVNGRGPDPGCLDZRLBGDJOZ/bundle.json","state_url":"https://pith.science/pith/42FNTVNGRGPDPGCLDZRLBGDJOZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/42FNTVNGRGPDPGCLDZRLBGDJOZ/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":"2026-06-10T12:00:57Z","links":{"resolver":"https://pith.science/pith/42FNTVNGRGPDPGCLDZRLBGDJOZ","bundle":"https://pith.science/pith/42FNTVNGRGPDPGCLDZRLBGDJOZ/bundle.json","state":"https://pith.science/pith/42FNTVNGRGPDPGCLDZRLBGDJOZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/42FNTVNGRGPDPGCLDZRLBGDJOZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2022:42FNTVNGRGPDPGCLDZRLBGDJOZ","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":"5c730c8b906cbb6fa41ef63a66f39d68103c6ce78288ef7e6e78d63b08517e89","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-07-28T17:40:47Z","title_canon_sha256":"e3f17dec2398ee2bed0424c1881217774b214a1303ffa79fbf895246d41b0e20"},"schema_version":"1.0","source":{"id":"2207.14255","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2207.14255","created_at":"2026-05-18T00:35:37Z"},{"alias_kind":"arxiv_version","alias_value":"2207.14255v1","created_at":"2026-05-18T00:35:37Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2207.14255","created_at":"2026-05-18T00:35:37Z"},{"alias_kind":"pith_short_12","alias_value":"42FNTVNGRGPD","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_16","alias_value":"42FNTVNGRGPDPGCL","created_at":"2026-05-18T12:33:33Z"},{"alias_kind":"pith_short_8","alias_value":"42FNTVNG","created_at":"2026-05-18T12:33:33Z"}],"graph_snapshots":[{"event_id":"sha256:352c7db04920b43f2d7e565520e264cd766c565e56badc4c4d89f8b2a9a5868d","target":"graph","created_at":"2026-05-18T00:35:37Z","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"},"paper":{"abstract_excerpt":"We show that autoregressive language models can learn to infill text after we apply a straightforward transformation to the dataset, which simply moves a span of text from the middle of a document to its end. While this data augmentation has garnered much interest in recent years, we provide extensive evidence that training models with a large fraction of data transformed in this way does not harm the original left-to-right generative capability, as measured by perplexity and sampling evaluations across a wide range of scales. Given the usefulness, simplicity, and efficiency of training models","authors_text":"Christine McLeavey, Heewoo Jun, Jerry Tworek, John Schulman, Mark Chen, Mohammad Bavarian, Nikolas Tezak","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-07-28T17:40:47Z","title":"Efficient Training of Language Models to Fill in the Middle"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2207.14255","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:f79297fe33015ae3e4987d69b3f797f061d8d312ff2572b1fc777c5c875f6f9c","target":"record","created_at":"2026-05-18T00:35:37Z","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":"5c730c8b906cbb6fa41ef63a66f39d68103c6ce78288ef7e6e78d63b08517e89","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-07-28T17:40:47Z","title_canon_sha256":"e3f17dec2398ee2bed0424c1881217774b214a1303ffa79fbf895246d41b0e20"},"schema_version":"1.0","source":{"id":"2207.14255","kind":"arxiv","version":1}},"canonical_sha256":"e68ad9d5a6899e37984b1e62b098697672f8bfa48de329b4fb0bf6ca0b714960","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e68ad9d5a6899e37984b1e62b098697672f8bfa48de329b4fb0bf6ca0b714960","first_computed_at":"2026-05-18T00:35:37.704002Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T00:35:37.704002Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"UooRznMuMzlmbRmNk3htWg0hBm1b4o5py8Gb+xffuP/IkwGxj9Gh+/U1Px0jD/L6qf6SjCvIhg78XKpVzIEuCA==","signature_status":"signed_v1","signed_at":"2026-05-18T00:35:37.704631Z","signed_message":"canonical_sha256_bytes"},"source_id":"2207.14255","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f79297fe33015ae3e4987d69b3f797f061d8d312ff2572b1fc777c5c875f6f9c","sha256:352c7db04920b43f2d7e565520e264cd766c565e56badc4c4d89f8b2a9a5868d"],"state_sha256":"7c758e234495faf11e5d7225fb0cd5b92419e809ec2bac8e502a67a0c4571ff0"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"g6lMyhOBEs0VItedMGE/juJJkSYgHdsIny7f20eusPI38Vs8xfwbk30jM3sbB+fc8l7wm8FyT8Iowu9ZnLxgBQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-10T12:00:57.520582Z","bundle_sha256":"335158133f8b9f7bfe32a0efd0f070a5a8f7349b201cc2d3856dc0d1529da15c"}}