{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:RFHR7JY6KFAWP6JJVEOL7YFBKA","short_pith_number":"pith:RFHR7JY6","schema_version":"1.0","canonical_sha256":"894f1fa71e514167f929a91cbfe0a1502a1e54d723ebe6efcb9d6f2f525cb3ca","source":{"kind":"arxiv","id":"2208.11857","version":2},"attestation_state":"computed","paper":{"title":"Shortcut Learning of Large Language Models in Natural Language Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Dacheng Tao, Fengxiang He, Mengnan Du, Na Zou, Xia Hu","submitted_at":"2022-08-25T03:51:39Z","abstract_excerpt":"Large language models (LLMs) have achieved state-of-the-art performance on a series of natural language understanding tasks. However, these LLMs might rely on dataset bias and artifacts as shortcuts for prediction. This has significantly affected their generalizability and adversarial robustness. In this paper, we provide a review of recent developments that address the shortcut learning and robustness challenge of LLMs. We first introduce the concepts of shortcut learning of language models. We then introduce methods to identify shortcut learning behavior in language models, characterize the "},"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":"2208.11857","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-08-25T03:51:39Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"be97950ae5cb00b82089e6a88beb8eb9e1b8307cb7a2b4ab0cf66ea31147f324","abstract_canon_sha256":"4966a78944deab6f857176eab0651d7825d6c6189619aa59ba12b0e370a68cc2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:07:32.234699Z","signature_b64":"3mHMpRk+gUu52m5uympqo64WOWwn5lacVC/KQnXz1SvY6jMtFMbvSZrJeZWKGFIbI4Td3/rffrDFXDbSc+XICQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"894f1fa71e514167f929a91cbfe0a1502a1e54d723ebe6efcb9d6f2f525cb3ca","last_reissued_at":"2026-07-05T06:07:32.234303Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:07:32.234303Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Shortcut Learning of Large Language Models in Natural Language Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Dacheng Tao, Fengxiang He, Mengnan Du, Na Zou, Xia Hu","submitted_at":"2022-08-25T03:51:39Z","abstract_excerpt":"Large language models (LLMs) have achieved state-of-the-art performance on a series of natural language understanding tasks. However, these LLMs might rely on dataset bias and artifacts as shortcuts for prediction. This has significantly affected their generalizability and adversarial robustness. In this paper, we provide a review of recent developments that address the shortcut learning and robustness challenge of LLMs. We first introduce the concepts of shortcut learning of language models. We then introduce methods to identify shortcut learning behavior in language models, characterize the "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2208.11857","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/2208.11857/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":"2208.11857","created_at":"2026-07-05T06:07:32.234368+00:00"},{"alias_kind":"arxiv_version","alias_value":"2208.11857v2","created_at":"2026-07-05T06:07:32.234368+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2208.11857","created_at":"2026-07-05T06:07:32.234368+00:00"},{"alias_kind":"pith_short_12","alias_value":"RFHR7JY6KFAW","created_at":"2026-07-05T06:07:32.234368+00:00"},{"alias_kind":"pith_short_16","alias_value":"RFHR7JY6KFAWP6JJ","created_at":"2026-07-05T06:07:32.234368+00:00"},{"alias_kind":"pith_short_8","alias_value":"RFHR7JY6","created_at":"2026-07-05T06:07:32.234368+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.30462","citing_title":"idSCD: Identifying Training Datasets through Semantic Correlation Descriptors","ref_index":11,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RFHR7JY6KFAWP6JJVEOL7YFBKA","json":"https://pith.science/pith/RFHR7JY6KFAWP6JJVEOL7YFBKA.json","graph_json":"https://pith.science/api/pith-number/RFHR7JY6KFAWP6JJVEOL7YFBKA/graph.json","events_json":"https://pith.science/api/pith-number/RFHR7JY6KFAWP6JJVEOL7YFBKA/events.json","paper":"https://pith.science/paper/RFHR7JY6"},"agent_actions":{"view_html":"https://pith.science/pith/RFHR7JY6KFAWP6JJVEOL7YFBKA","download_json":"https://pith.science/pith/RFHR7JY6KFAWP6JJVEOL7YFBKA.json","view_paper":"https://pith.science/paper/RFHR7JY6","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2208.11857&json=true","fetch_graph":"https://pith.science/api/pith-number/RFHR7JY6KFAWP6JJVEOL7YFBKA/graph.json","fetch_events":"https://pith.science/api/pith-number/RFHR7JY6KFAWP6JJVEOL7YFBKA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RFHR7JY6KFAWP6JJVEOL7YFBKA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RFHR7JY6KFAWP6JJVEOL7YFBKA/action/storage_attestation","attest_author":"https://pith.science/pith/RFHR7JY6KFAWP6JJVEOL7YFBKA/action/author_attestation","sign_citation":"https://pith.science/pith/RFHR7JY6KFAWP6JJVEOL7YFBKA/action/citation_signature","submit_replication":"https://pith.science/pith/RFHR7JY6KFAWP6JJVEOL7YFBKA/action/replication_record"}},"created_at":"2026-07-05T06:07:32.234368+00:00","updated_at":"2026-07-05T06:07:32.234368+00:00"}