{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2021:SR3QPI27VRY2TST4E3Z2GO64U5","short_pith_number":"pith:SR3QPI27","canonical_record":{"source":{"id":"2103.00453","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-02-28T11:07:37Z","cross_cats_sorted":[],"title_canon_sha256":"de5781f1c38cc863cacba016452797a90f633536c714a0ca2a9f9e8354d51d7a","abstract_canon_sha256":"1d35e85e5ed5788398f25e69efbd14d326f06c9ef1ce9dc8d4db85f13ccb7dee"},"schema_version":"1.0"},"canonical_sha256":"947707a35fac71a9ca7c26f3a33bdca755a0f5d8c1d3cb4652252481736d2fb9","source":{"kind":"arxiv","id":"2103.00453","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2103.00453","created_at":"2026-07-05T03:12:47Z"},{"alias_kind":"arxiv_version","alias_value":"2103.00453v2","created_at":"2026-07-05T03:12:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.00453","created_at":"2026-07-05T03:12:47Z"},{"alias_kind":"pith_short_12","alias_value":"SR3QPI27VRY2","created_at":"2026-07-05T03:12:47Z"},{"alias_kind":"pith_short_16","alias_value":"SR3QPI27VRY2TST4","created_at":"2026-07-05T03:12:47Z"},{"alias_kind":"pith_short_8","alias_value":"SR3QPI27","created_at":"2026-07-05T03:12:47Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2021:SR3QPI27VRY2TST4E3Z2GO64U5","target":"record","payload":{"canonical_record":{"source":{"id":"2103.00453","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-02-28T11:07:37Z","cross_cats_sorted":[],"title_canon_sha256":"de5781f1c38cc863cacba016452797a90f633536c714a0ca2a9f9e8354d51d7a","abstract_canon_sha256":"1d35e85e5ed5788398f25e69efbd14d326f06c9ef1ce9dc8d4db85f13ccb7dee"},"schema_version":"1.0"},"canonical_sha256":"947707a35fac71a9ca7c26f3a33bdca755a0f5d8c1d3cb4652252481736d2fb9","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:12:47.024286Z","signature_b64":"iq9x6K5mP/Auv2SmdhiJ4Juf1zm3T3itFDmEv4KJqiCQIHSsnQnL1bpCKRTo1cAK8TRc08TK1AR7Dlhe0H6NBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"947707a35fac71a9ca7c26f3a33bdca755a0f5d8c1d3cb4652252481736d2fb9","last_reissued_at":"2026-07-05T03:12:47.023789Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:12:47.023789Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2103.00453","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":"2026-07-05T03:12:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jQYZ3LFqrAmQsbnEXF7Z7Q48j7bZYl3E63xKhs6JhB51jzxUOQAEZU4vzdbVYcPV94RbJF4xII0lbUuzI+vqAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T09:18:20.884549Z"},"content_sha256":"aa8f7cc3dc1dd3f3504f8ee4ef4d5ce75343dca9f9117e41bb5a5916f9129796","schema_version":"1.0","event_id":"sha256:aa8f7cc3dc1dd3f3504f8ee4ef4d5ce75343dca9f9117e41bb5a5916f9129796"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2021:SR3QPI27VRY2TST4E3Z2GO64U5","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hinrich Sch\\\"utze, Sahana Udupa, Timo Schick","submitted_at":"2021-02-28T11:07:37Z","abstract_excerpt":"When trained on large, unfiltered crawls from the internet, language models pick up and reproduce all kinds of undesirable biases that can be found in the data: they often generate racist, sexist, violent or otherwise toxic language. As large models require millions of training examples to achieve good performance, it is difficult to completely prevent them from being exposed to such content. In this paper, we first demonstrate a surprising finding: pretrained language models recognize, to a considerable degree, their undesirable biases and the toxicity of the content they produce. We refer to"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.00453","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/2103.00453/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":"2026-07-05T03:12:47Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZKXxISZFgnLRhmqXoE0LvIM14ZlYxgiyy9nfi4U0ovqOgK/y6kUeM3D8dovujaJUXhwK1SkCC4uiGG81e5/sCQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-07T09:18:20.885063Z"},"content_sha256":"38f9895a4b609699ba71273f08e3ea26126e694b51de7cac217707d5ae7c58da","schema_version":"1.0","event_id":"sha256:38f9895a4b609699ba71273f08e3ea26126e694b51de7cac217707d5ae7c58da"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/SR3QPI27VRY2TST4E3Z2GO64U5/bundle.json","state_url":"https://pith.science/pith/SR3QPI27VRY2TST4E3Z2GO64U5/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/SR3QPI27VRY2TST4E3Z2GO64U5/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":"2026-08-07T09:18:20Z","links":{"resolver":"https://pith.science/pith/SR3QPI27VRY2TST4E3Z2GO64U5","bundle":"https://pith.science/pith/SR3QPI27VRY2TST4E3Z2GO64U5/bundle.json","state":"https://pith.science/pith/SR3QPI27VRY2TST4E3Z2GO64U5/state.json","well_known_bundle":"https://pith.science/.well-known/pith/SR3QPI27VRY2TST4E3Z2GO64U5/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2021:SR3QPI27VRY2TST4E3Z2GO64U5","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":"1d35e85e5ed5788398f25e69efbd14d326f06c9ef1ce9dc8d4db85f13ccb7dee","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-02-28T11:07:37Z","title_canon_sha256":"de5781f1c38cc863cacba016452797a90f633536c714a0ca2a9f9e8354d51d7a"},"schema_version":"1.0","source":{"id":"2103.00453","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2103.00453","created_at":"2026-07-05T03:12:47Z"},{"alias_kind":"arxiv_version","alias_value":"2103.00453v2","created_at":"2026-07-05T03:12:47Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2103.00453","created_at":"2026-07-05T03:12:47Z"},{"alias_kind":"pith_short_12","alias_value":"SR3QPI27VRY2","created_at":"2026-07-05T03:12:47Z"},{"alias_kind":"pith_short_16","alias_value":"SR3QPI27VRY2TST4","created_at":"2026-07-05T03:12:47Z"},{"alias_kind":"pith_short_8","alias_value":"SR3QPI27","created_at":"2026-07-05T03:12:47Z"}],"graph_snapshots":[{"event_id":"sha256:38f9895a4b609699ba71273f08e3ea26126e694b51de7cac217707d5ae7c58da","target":"graph","created_at":"2026-07-05T03:12:47Z","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/2103.00453/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"When trained on large, unfiltered crawls from the internet, language models pick up and reproduce all kinds of undesirable biases that can be found in the data: they often generate racist, sexist, violent or otherwise toxic language. As large models require millions of training examples to achieve good performance, it is difficult to completely prevent them from being exposed to such content. In this paper, we first demonstrate a surprising finding: pretrained language models recognize, to a considerable degree, their undesirable biases and the toxicity of the content they produce. We refer to","authors_text":"Hinrich Sch\\\"utze, Sahana Udupa, Timo Schick","cross_cats":[],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-02-28T11:07:37Z","title":"Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2103.00453","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:aa8f7cc3dc1dd3f3504f8ee4ef4d5ce75343dca9f9117e41bb5a5916f9129796","target":"record","created_at":"2026-07-05T03:12:47Z","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":"1d35e85e5ed5788398f25e69efbd14d326f06c9ef1ce9dc8d4db85f13ccb7dee","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2021-02-28T11:07:37Z","title_canon_sha256":"de5781f1c38cc863cacba016452797a90f633536c714a0ca2a9f9e8354d51d7a"},"schema_version":"1.0","source":{"id":"2103.00453","kind":"arxiv","version":2}},"canonical_sha256":"947707a35fac71a9ca7c26f3a33bdca755a0f5d8c1d3cb4652252481736d2fb9","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"947707a35fac71a9ca7c26f3a33bdca755a0f5d8c1d3cb4652252481736d2fb9","first_computed_at":"2026-07-05T03:12:47.023789Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T03:12:47.023789Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"iq9x6K5mP/Auv2SmdhiJ4Juf1zm3T3itFDmEv4KJqiCQIHSsnQnL1bpCKRTo1cAK8TRc08TK1AR7Dlhe0H6NBA==","signature_status":"signed_v1","signed_at":"2026-07-05T03:12:47.024286Z","signed_message":"canonical_sha256_bytes"},"source_id":"2103.00453","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:aa8f7cc3dc1dd3f3504f8ee4ef4d5ce75343dca9f9117e41bb5a5916f9129796","sha256:38f9895a4b609699ba71273f08e3ea26126e694b51de7cac217707d5ae7c58da"],"state_sha256":"19fcffc476563c2d412220360b3fc00bd3d438881341c99acba38749464b4c80"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"VhoG+l+AGgCfrdNkxGpi8oawmjbY+GiiPDwpm/HamR9WDOSMKiRX54Xb8NI92j2cuv56MrKtuddFocIxWf68AQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-07T09:18:20.889300Z","bundle_sha256":"e62b8e58c4c9bbb3adbe035bb532f48cd22842f75505a0c6749c96c4588afaf2"}}