{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:FVKYRRHMDISESRAA53YXSTT6H4","short_pith_number":"pith:FVKYRRHM","schema_version":"1.0","canonical_sha256":"2d5588c4ec1a24494400eef1794e7e3f3e423adb1cfcd784db8fc1539343d311","source":{"kind":"arxiv","id":"2402.08225","version":3},"attestation_state":"computed","paper":{"title":"Improving Black-box Robustness with In-Context Rewriting","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hamid Palangi, Isha Puri, Jorge Mendez, Kyle O'Brien, Marzyeh Ghassemi, Nathan Ng, Thomas Hartvigsen, Yoon Kim","submitted_at":"2024-02-13T05:33:35Z","abstract_excerpt":"Machine learning models for text classification often excel on in-distribution (ID) data but struggle with unseen out-of-distribution (OOD) inputs. Most techniques for improving OOD robustness are not applicable to settings where the model is effectively a black box, such as when the weights are frozen, retraining is costly, or the model is leveraged via an API. Test-time augmentation (TTA) is a simple post-hoc technique for improving robustness that sidesteps black-box constraints by aggregating predictions across multiple augmentations of the test input. TTA has seen limited use in NLP due t"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2402.08225","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-02-13T05:33:35Z","cross_cats_sorted":[],"title_canon_sha256":"6a8b058c485957964ffe80791be567f295f674885d552f561387271812f67328","abstract_canon_sha256":"b91206b5d8fb490bc95167cb5d4e14b2d0df898cebe6d5c739a79c333bf0d08b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:51:47.190122Z","signature_b64":"yqf+yk/3gwcDzFVwAjVjt6jEFhx4UHxS2bWsVMYFDEj6AOV5HVb3fPVDRZ4NzTO0+lgtZCdKxA45KzYchWgbBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2d5588c4ec1a24494400eef1794e7e3f3e423adb1cfcd784db8fc1539343d311","last_reissued_at":"2026-07-05T08:51:47.189670Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:51:47.189670Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improving Black-box Robustness with In-Context Rewriting","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hamid Palangi, Isha Puri, Jorge Mendez, Kyle O'Brien, Marzyeh Ghassemi, Nathan Ng, Thomas Hartvigsen, Yoon Kim","submitted_at":"2024-02-13T05:33:35Z","abstract_excerpt":"Machine learning models for text classification often excel on in-distribution (ID) data but struggle with unseen out-of-distribution (OOD) inputs. Most techniques for improving OOD robustness are not applicable to settings where the model is effectively a black box, such as when the weights are frozen, retraining is costly, or the model is leveraged via an API. Test-time augmentation (TTA) is a simple post-hoc technique for improving robustness that sidesteps black-box constraints by aggregating predictions across multiple augmentations of the test input. TTA has seen limited use in NLP due t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.08225","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/2402.08225/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2402.08225","created_at":"2026-07-05T08:51:47.189728+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.08225v3","created_at":"2026-07-05T08:51:47.189728+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.08225","created_at":"2026-07-05T08:51:47.189728+00:00"},{"alias_kind":"pith_short_12","alias_value":"FVKYRRHMDISE","created_at":"2026-07-05T08:51:47.189728+00:00"},{"alias_kind":"pith_short_16","alias_value":"FVKYRRHMDISESRAA","created_at":"2026-07-05T08:51:47.189728+00:00"},{"alias_kind":"pith_short_8","alias_value":"FVKYRRHM","created_at":"2026-07-05T08:51:47.189728+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.10535","citing_title":"On Adversarial Robustness and Out-of-Distribution Robustness of Large Language Models","ref_index":5,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/FVKYRRHMDISESRAA53YXSTT6H4","json":"https://pith.science/pith/FVKYRRHMDISESRAA53YXSTT6H4.json","graph_json":"https://pith.science/api/pith-number/FVKYRRHMDISESRAA53YXSTT6H4/graph.json","events_json":"https://pith.science/api/pith-number/FVKYRRHMDISESRAA53YXSTT6H4/events.json","paper":"https://pith.science/paper/FVKYRRHM"},"agent_actions":{"view_html":"https://pith.science/pith/FVKYRRHMDISESRAA53YXSTT6H4","download_json":"https://pith.science/pith/FVKYRRHMDISESRAA53YXSTT6H4.json","view_paper":"https://pith.science/paper/FVKYRRHM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.08225&json=true","fetch_graph":"https://pith.science/api/pith-number/FVKYRRHMDISESRAA53YXSTT6H4/graph.json","fetch_events":"https://pith.science/api/pith-number/FVKYRRHMDISESRAA53YXSTT6H4/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FVKYRRHMDISESRAA53YXSTT6H4/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FVKYRRHMDISESRAA53YXSTT6H4/action/storage_attestation","attest_author":"https://pith.science/pith/FVKYRRHMDISESRAA53YXSTT6H4/action/author_attestation","sign_citation":"https://pith.science/pith/FVKYRRHMDISESRAA53YXSTT6H4/action/citation_signature","submit_replication":"https://pith.science/pith/FVKYRRHMDISESRAA53YXSTT6H4/action/replication_record"}},"created_at":"2026-07-05T08:51:47.189728+00:00","updated_at":"2026-07-05T08:51:47.189728+00:00"}