{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2019:HIGE7GOH32K5UEJBZFWD5HNJ34","short_pith_number":"pith:HIGE7GOH","canonical_record":{"source":{"id":"1910.14659","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-10-31T17:51:21Z","cross_cats_sorted":["cs.LG","eess.AS","stat.ML"],"title_canon_sha256":"a96dd03b27e0f4d65a67935d17121ef7ab955503be06b02463c84d2f0c97f009","abstract_canon_sha256":"421881c0ada7895d55056af941c5cbb2b004371c2ac781264ca479d0b7d5d779"},"schema_version":"1.0"},"canonical_sha256":"3a0c4f99c7de95da1121c96c3e9da9df2d2a7c13668e034f9ee63362e3f84519","source":{"kind":"arxiv","id":"1910.14659","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1910.14659","created_at":"2026-07-05T02:04:04Z"},{"alias_kind":"arxiv_version","alias_value":"1910.14659v3","created_at":"2026-07-05T02:04:04Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1910.14659","created_at":"2026-07-05T02:04:04Z"},{"alias_kind":"pith_short_12","alias_value":"HIGE7GOH32K5","created_at":"2026-07-05T02:04:04Z"},{"alias_kind":"pith_short_16","alias_value":"HIGE7GOH32K5UEJB","created_at":"2026-07-05T02:04:04Z"},{"alias_kind":"pith_short_8","alias_value":"HIGE7GOH","created_at":"2026-07-05T02:04:04Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2019:HIGE7GOH32K5UEJBZFWD5HNJ34","target":"record","payload":{"canonical_record":{"source":{"id":"1910.14659","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-10-31T17:51:21Z","cross_cats_sorted":["cs.LG","eess.AS","stat.ML"],"title_canon_sha256":"a96dd03b27e0f4d65a67935d17121ef7ab955503be06b02463c84d2f0c97f009","abstract_canon_sha256":"421881c0ada7895d55056af941c5cbb2b004371c2ac781264ca479d0b7d5d779"},"schema_version":"1.0"},"canonical_sha256":"3a0c4f99c7de95da1121c96c3e9da9df2d2a7c13668e034f9ee63362e3f84519","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:04:04.500266Z","signature_b64":"FTY5KnXOHqhYDh1KDrTIFgsyRBaHoOHJf7KA56OVD9tKMbKAp1XMoRe5bpD8ZtDKYSS7FPoMJmgnlLoQXrOjDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3a0c4f99c7de95da1121c96c3e9da9df2d2a7c13668e034f9ee63362e3f84519","last_reissued_at":"2026-07-05T02:04:04.499860Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:04:04.499860Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1910.14659","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":"2026-07-05T02:04:04Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"DqedqPGkW/LU7WQEIL9j4PGc1rymHiXR7LGr83IdwNKb/1pg7ENAOrlV5GjsCeZhAaMZ43XnrBHiLfa6PJcDAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T01:36:21.178669Z"},"content_sha256":"9d907b6b4d3ef8918eef0eb505c02cc434d44036195722e991fda7aaafeda012","schema_version":"1.0","event_id":"sha256:9d907b6b4d3ef8918eef0eb505c02cc434d44036195722e991fda7aaafeda012"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2019:HIGE7GOH32K5UEJBZFWD5HNJ34","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Masked Language Model Scoring","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","eess.AS","stat.ML"],"primary_cat":"cs.CL","authors_text":"Davis Liang, Julian Salazar, Katrin Kirchhoff, Toan Q. Nguyen","submitted_at":"2019-10-31T17:51:21Z","abstract_excerpt":"Pretrained masked language models (MLMs) require finetuning for most NLP tasks. Instead, we evaluate MLMs out of the box via their pseudo-log-likelihood scores (PLLs), which are computed by masking tokens one by one. We show that PLLs outperform scores from autoregressive language models like GPT-2 in a variety of tasks. By rescoring ASR and NMT hypotheses, RoBERTa reduces an end-to-end LibriSpeech model's WER by 30% relative and adds up to +1.7 BLEU on state-of-the-art baselines for low-resource translation pairs, with further gains from domain adaptation. We attribute this success to PLL's u"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1910.14659","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/1910.14659/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":"2026-07-05T02:04:04Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dlRZXQJUY/NXeNUGDVT6WB07sRw+NuzhGHCO9lJFDMFrZ0bfyV3wS/BTgjRZ5i/L8GSrxx1l1Kd5/WWripjNAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T01:36:21.179162Z"},"content_sha256":"58416d2a69c2c34e682640db1263b7d56d7d52d2b2ee3a93ca8dd5de735f1512","schema_version":"1.0","event_id":"sha256:58416d2a69c2c34e682640db1263b7d56d7d52d2b2ee3a93ca8dd5de735f1512"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/HIGE7GOH32K5UEJBZFWD5HNJ34/bundle.json","state_url":"https://pith.science/pith/HIGE7GOH32K5UEJBZFWD5HNJ34/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/HIGE7GOH32K5UEJBZFWD5HNJ34/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":"2026-08-09T01:36:21Z","links":{"resolver":"https://pith.science/pith/HIGE7GOH32K5UEJBZFWD5HNJ34","bundle":"https://pith.science/pith/HIGE7GOH32K5UEJBZFWD5HNJ34/bundle.json","state":"https://pith.science/pith/HIGE7GOH32K5UEJBZFWD5HNJ34/state.json","well_known_bundle":"https://pith.science/.well-known/pith/HIGE7GOH32K5UEJBZFWD5HNJ34/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2019:HIGE7GOH32K5UEJBZFWD5HNJ34","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":"421881c0ada7895d55056af941c5cbb2b004371c2ac781264ca479d0b7d5d779","cross_cats_sorted":["cs.LG","eess.AS","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-10-31T17:51:21Z","title_canon_sha256":"a96dd03b27e0f4d65a67935d17121ef7ab955503be06b02463c84d2f0c97f009"},"schema_version":"1.0","source":{"id":"1910.14659","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1910.14659","created_at":"2026-07-05T02:04:04Z"},{"alias_kind":"arxiv_version","alias_value":"1910.14659v3","created_at":"2026-07-05T02:04:04Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1910.14659","created_at":"2026-07-05T02:04:04Z"},{"alias_kind":"pith_short_12","alias_value":"HIGE7GOH32K5","created_at":"2026-07-05T02:04:04Z"},{"alias_kind":"pith_short_16","alias_value":"HIGE7GOH32K5UEJB","created_at":"2026-07-05T02:04:04Z"},{"alias_kind":"pith_short_8","alias_value":"HIGE7GOH","created_at":"2026-07-05T02:04:04Z"}],"graph_snapshots":[{"event_id":"sha256:58416d2a69c2c34e682640db1263b7d56d7d52d2b2ee3a93ca8dd5de735f1512","target":"graph","created_at":"2026-07-05T02:04:04Z","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/1910.14659/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Pretrained masked language models (MLMs) require finetuning for most NLP tasks. Instead, we evaluate MLMs out of the box via their pseudo-log-likelihood scores (PLLs), which are computed by masking tokens one by one. We show that PLLs outperform scores from autoregressive language models like GPT-2 in a variety of tasks. By rescoring ASR and NMT hypotheses, RoBERTa reduces an end-to-end LibriSpeech model's WER by 30% relative and adds up to +1.7 BLEU on state-of-the-art baselines for low-resource translation pairs, with further gains from domain adaptation. We attribute this success to PLL's u","authors_text":"Davis Liang, Julian Salazar, Katrin Kirchhoff, Toan Q. Nguyen","cross_cats":["cs.LG","eess.AS","stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-10-31T17:51:21Z","title":"Masked Language Model Scoring"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1910.14659","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":"sha256:9d907b6b4d3ef8918eef0eb505c02cc434d44036195722e991fda7aaafeda012","target":"record","created_at":"2026-07-05T02:04:04Z","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":"421881c0ada7895d55056af941c5cbb2b004371c2ac781264ca479d0b7d5d779","cross_cats_sorted":["cs.LG","eess.AS","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2019-10-31T17:51:21Z","title_canon_sha256":"a96dd03b27e0f4d65a67935d17121ef7ab955503be06b02463c84d2f0c97f009"},"schema_version":"1.0","source":{"id":"1910.14659","kind":"arxiv","version":3}},"canonical_sha256":"3a0c4f99c7de95da1121c96c3e9da9df2d2a7c13668e034f9ee63362e3f84519","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3a0c4f99c7de95da1121c96c3e9da9df2d2a7c13668e034f9ee63362e3f84519","first_computed_at":"2026-07-05T02:04:04.499860Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T02:04:04.499860Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"FTY5KnXOHqhYDh1KDrTIFgsyRBaHoOHJf7KA56OVD9tKMbKAp1XMoRe5bpD8ZtDKYSS7FPoMJmgnlLoQXrOjDw==","signature_status":"signed_v1","signed_at":"2026-07-05T02:04:04.500266Z","signed_message":"canonical_sha256_bytes"},"source_id":"1910.14659","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9d907b6b4d3ef8918eef0eb505c02cc434d44036195722e991fda7aaafeda012","sha256:58416d2a69c2c34e682640db1263b7d56d7d52d2b2ee3a93ca8dd5de735f1512"],"state_sha256":"ff3e3d2adf34ddc19c3910366f0c51fd9482891111d78ac764988a6df39f41f7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"GYMKyzRZf3fWNauiQKxNZR4DO/iCzs8IWoKX0MEO4bjtIejS1FcCuCBnjTmlAWlwvVpHnSUoZC7rKKeyQSPeCA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T01:36:21.182509Z","bundle_sha256":"43ff0ffa641a2d02724187f95cd36fac359ca25d118b5c1ecca20536b0a93550"}}