{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2021:NUW4Z3TJDBFXVGXS2T3MYOYCZZ","short_pith_number":"pith:NUW4Z3TJ","canonical_record":{"source":{"id":"2101.00288","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2021-01-01T18:34:22Z","cross_cats_sorted":[],"title_canon_sha256":"37c8f635ba604517fb0cc05ded68253a35fc5f41bf9d3257cf80559efce57c50","abstract_canon_sha256":"c933015334c45c7a3fc24f9b7f3863864b60fed4f21a969ca7193078c6530980"},"schema_version":"1.0"},"canonical_sha256":"6d2dccee69184b7a9af2d4f6cc3b02ce447b77037d5b6dd1ff3107274ac64a26","source":{"kind":"arxiv","id":"2101.00288","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2101.00288","created_at":"2026-07-05T02:44:53Z"},{"alias_kind":"arxiv_version","alias_value":"2101.00288v2","created_at":"2026-07-05T02:44:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2101.00288","created_at":"2026-07-05T02:44:53Z"},{"alias_kind":"pith_short_12","alias_value":"NUW4Z3TJDBFX","created_at":"2026-07-05T02:44:53Z"},{"alias_kind":"pith_short_16","alias_value":"NUW4Z3TJDBFXVGXS","created_at":"2026-07-05T02:44:53Z"},{"alias_kind":"pith_short_8","alias_value":"NUW4Z3TJ","created_at":"2026-07-05T02:44:53Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2021:NUW4Z3TJDBFXVGXS2T3MYOYCZZ","target":"record","payload":{"canonical_record":{"source":{"id":"2101.00288","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2021-01-01T18:34:22Z","cross_cats_sorted":[],"title_canon_sha256":"37c8f635ba604517fb0cc05ded68253a35fc5f41bf9d3257cf80559efce57c50","abstract_canon_sha256":"c933015334c45c7a3fc24f9b7f3863864b60fed4f21a969ca7193078c6530980"},"schema_version":"1.0"},"canonical_sha256":"6d2dccee69184b7a9af2d4f6cc3b02ce447b77037d5b6dd1ff3107274ac64a26","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:44:53.305362Z","signature_b64":"vLvxEocG9kqoB9XDhVYgLk8lbrJasLsUU510I52LKFWjlFrWzXKwtVByOiryWabOX8droE8Rtk5XuGAD/FRyCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6d2dccee69184b7a9af2d4f6cc3b02ce447b77037d5b6dd1ff3107274ac64a26","last_reissued_at":"2026-07-05T02:44:53.304741Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:44:53.304741Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2101.00288","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":"2026-07-05T02:44:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ZnGVB8fZvd7LM9l+FQ17AHQ7w4ar1irnzKU5SCwhwoeyXP5LTb4EwlVSDxmLqt9vlRizXkU0ACXSM2Yyp8UDDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T16:58:36.801489Z"},"content_sha256":"0c6d8633a2588d2396b95fb250255e87071c59584c2021d458f052dde600990d","schema_version":"1.0","event_id":"sha256:0c6d8633a2588d2396b95fb250255e87071c59584c2021d458f052dde600990d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2021:NUW4Z3TJDBFXVGXS2T3MYOYCZZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Polyjuice: Generating Counterfactuals for Explaining, Evaluating, and Improving Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Daniel S. Weld, Jeffrey Heer, Marco Tulio Ribeiro, Tongshuang Wu","submitted_at":"2021-01-01T18:34:22Z","abstract_excerpt":"While counterfactual examples are useful for analysis and training of NLP models, current generation methods either rely on manual labor to create very few counterfactuals, or only instantiate limited types of perturbations such as paraphrases or word substitutions. We present Polyjuice, a general-purpose counterfactual generator that allows for control over perturbation types and locations, trained by finetuning GPT-2 on multiple datasets of paired sentences. We show that Polyjuice produces diverse sets of realistic counterfactuals, which in turn are useful in various distinct applications: i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2101.00288","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/2101.00288/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":"2026-07-05T02:44:53Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"heHtdR3XQSyS2DgBEXbXI7RyQ3vPVoM6NKIJ+sEtbPw4ylg8gg79R/5A9ozfCss8MPYnDUAU5gGylyKMWB5rDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T16:58:36.801986Z"},"content_sha256":"fcd42c35797d872cc0e0d5e043e52168fb88fd5533a2dcf2d56f6bf7f4da2bf2","schema_version":"1.0","event_id":"sha256:fcd42c35797d872cc0e0d5e043e52168fb88fd5533a2dcf2d56f6bf7f4da2bf2"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/NUW4Z3TJDBFXVGXS2T3MYOYCZZ/bundle.json","state_url":"https://pith.science/pith/NUW4Z3TJDBFXVGXS2T3MYOYCZZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/NUW4Z3TJDBFXVGXS2T3MYOYCZZ/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":"2026-08-08T16:58:36Z","links":{"resolver":"https://pith.science/pith/NUW4Z3TJDBFXVGXS2T3MYOYCZZ","bundle":"https://pith.science/pith/NUW4Z3TJDBFXVGXS2T3MYOYCZZ/bundle.json","state":"https://pith.science/pith/NUW4Z3TJDBFXVGXS2T3MYOYCZZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/NUW4Z3TJDBFXVGXS2T3MYOYCZZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2021:NUW4Z3TJDBFXVGXS2T3MYOYCZZ","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":"c933015334c45c7a3fc24f9b7f3863864b60fed4f21a969ca7193078c6530980","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2021-01-01T18:34:22Z","title_canon_sha256":"37c8f635ba604517fb0cc05ded68253a35fc5f41bf9d3257cf80559efce57c50"},"schema_version":"1.0","source":{"id":"2101.00288","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2101.00288","created_at":"2026-07-05T02:44:53Z"},{"alias_kind":"arxiv_version","alias_value":"2101.00288v2","created_at":"2026-07-05T02:44:53Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2101.00288","created_at":"2026-07-05T02:44:53Z"},{"alias_kind":"pith_short_12","alias_value":"NUW4Z3TJDBFX","created_at":"2026-07-05T02:44:53Z"},{"alias_kind":"pith_short_16","alias_value":"NUW4Z3TJDBFXVGXS","created_at":"2026-07-05T02:44:53Z"},{"alias_kind":"pith_short_8","alias_value":"NUW4Z3TJ","created_at":"2026-07-05T02:44:53Z"}],"graph_snapshots":[{"event_id":"sha256:fcd42c35797d872cc0e0d5e043e52168fb88fd5533a2dcf2d56f6bf7f4da2bf2","target":"graph","created_at":"2026-07-05T02:44:53Z","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/2101.00288/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"While counterfactual examples are useful for analysis and training of NLP models, current generation methods either rely on manual labor to create very few counterfactuals, or only instantiate limited types of perturbations such as paraphrases or word substitutions. We present Polyjuice, a general-purpose counterfactual generator that allows for control over perturbation types and locations, trained by finetuning GPT-2 on multiple datasets of paired sentences. We show that Polyjuice produces diverse sets of realistic counterfactuals, which in turn are useful in various distinct applications: i","authors_text":"Daniel S. Weld, Jeffrey Heer, Marco Tulio Ribeiro, Tongshuang Wu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2021-01-01T18:34:22Z","title":"Polyjuice: Generating Counterfactuals for Explaining, Evaluating, and Improving Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2101.00288","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":"sha256:0c6d8633a2588d2396b95fb250255e87071c59584c2021d458f052dde600990d","target":"record","created_at":"2026-07-05T02:44:53Z","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":"c933015334c45c7a3fc24f9b7f3863864b60fed4f21a969ca7193078c6530980","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2021-01-01T18:34:22Z","title_canon_sha256":"37c8f635ba604517fb0cc05ded68253a35fc5f41bf9d3257cf80559efce57c50"},"schema_version":"1.0","source":{"id":"2101.00288","kind":"arxiv","version":2}},"canonical_sha256":"6d2dccee69184b7a9af2d4f6cc3b02ce447b77037d5b6dd1ff3107274ac64a26","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6d2dccee69184b7a9af2d4f6cc3b02ce447b77037d5b6dd1ff3107274ac64a26","first_computed_at":"2026-07-05T02:44:53.304741Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T02:44:53.304741Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"vLvxEocG9kqoB9XDhVYgLk8lbrJasLsUU510I52LKFWjlFrWzXKwtVByOiryWabOX8droE8Rtk5XuGAD/FRyCg==","signature_status":"signed_v1","signed_at":"2026-07-05T02:44:53.305362Z","signed_message":"canonical_sha256_bytes"},"source_id":"2101.00288","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:0c6d8633a2588d2396b95fb250255e87071c59584c2021d458f052dde600990d","sha256:fcd42c35797d872cc0e0d5e043e52168fb88fd5533a2dcf2d56f6bf7f4da2bf2"],"state_sha256":"7e6eb2cbf9421b38e34b3f466e30f5159a9dce2d5f83372fe9e4800f29d3083b"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jiIi8rKXjbppTfTF6wAqQg2FqmMe4fK/c+Bn/JrLpPHizYRB3uTXdEBJor2Bmg3ersRZpfT7GCIWlLTOSOHrBA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T16:58:36.806279Z","bundle_sha256":"dd9f4cf518cb008d433ea422b0c49d3d9802f3f4873d13cc75a4b2691b3ec461"}}