{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2017:VBYQDTRMQU7BSLKVWGH2JITEMK","short_pith_number":"pith:VBYQDTRM","schema_version":"1.0","canonical_sha256":"a87101ce2c853e192d55b18fa4a264629f49ff90e324da18d8194535dda27ceb","source":{"kind":"arxiv","id":"1712.06751","version":2},"attestation_state":"computed","paper":{"title":"HotFlip: White-Box Adversarial Examples for Text Classification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Anyi Rao, Daniel Lowd, Dejing Dou, Javid Ebrahimi","submitted_at":"2017-12-19T02:15:19Z","abstract_excerpt":"We propose an efficient method to generate white-box adversarial examples to trick a character-level neural classifier. We find that only a few manipulations are needed to greatly decrease the accuracy. Our method relies on an atomic flip operation, which swaps one token for another, based on the gradients of the one-hot input vectors. Due to efficiency of our method, we can perform adversarial training which makes the model more robust to attacks at test time. With the use of a few semantics-preserving constraints, we demonstrate that HotFlip can be adapted to attack a word-level classifier a"},"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":"1712.06751","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2017-12-19T02:15:19Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"8163a56ecaf719d1bf592bcc823cc50f365dfdb5179d8a0c3f097e1d56844243","abstract_canon_sha256":"2dd2c489ce380d3123e974b565ac1573ad9826f42dd7c12cd6c297e7dbb93e0f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:15:05.144496Z","signature_b64":"RqqqelO1j9mbpop9OnYpjHKHyS+YgpPvACj41TAbmcCqqs03FytsuxLHZcDvSm0PEBXJH5rz3mE8n0cYQbc4BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a87101ce2c853e192d55b18fa4a264629f49ff90e324da18d8194535dda27ceb","last_reissued_at":"2026-05-18T00:15:05.143745Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:15:05.143745Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"HotFlip: White-Box Adversarial Examples for Text Classification","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CL","authors_text":"Anyi Rao, Daniel Lowd, Dejing Dou, Javid Ebrahimi","submitted_at":"2017-12-19T02:15:19Z","abstract_excerpt":"We propose an efficient method to generate white-box adversarial examples to trick a character-level neural classifier. We find that only a few manipulations are needed to greatly decrease the accuracy. Our method relies on an atomic flip operation, which swaps one token for another, based on the gradients of the one-hot input vectors. Due to efficiency of our method, we can perform adversarial training which makes the model more robust to attacks at test time. With the use of a few semantics-preserving constraints, we demonstrate that HotFlip can be adapted to attack a word-level classifier a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1712.06751","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":""},"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":"1712.06751","created_at":"2026-05-18T00:15:05.143872+00:00"},{"alias_kind":"arxiv_version","alias_value":"1712.06751v2","created_at":"2026-05-18T00:15:05.143872+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1712.06751","created_at":"2026-05-18T00:15:05.143872+00:00"},{"alias_kind":"pith_short_12","alias_value":"VBYQDTRMQU7B","created_at":"2026-05-18T12:31:49.984773+00:00"},{"alias_kind":"pith_short_16","alias_value":"VBYQDTRMQU7BSLKV","created_at":"2026-05-18T12:31:49.984773+00:00"},{"alias_kind":"pith_short_8","alias_value":"VBYQDTRM","created_at":"2026-05-18T12:31:49.984773+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":9,"internal_anchor_count":3,"sample":[{"citing_arxiv_id":"2307.15043","citing_title":"Universal and Transferable Adversarial Attacks on Aligned Language Models","ref_index":6,"is_internal_anchor":true},{"citing_arxiv_id":"2508.03949","citing_title":"Model Compression vs. Adversarial Robustness: An Empirical Study on Language Models for Code","ref_index":54,"is_internal_anchor":true},{"citing_arxiv_id":"2504.19793","citing_title":"Prompt Injection Attack to Tool Selection in LLM Agents","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2604.28157","citing_title":"FlashRT: Towards Computationally and Memory Efficient Red-Teaming for Prompt Injection and Knowledge Corruption","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03482","citing_title":"MEMSAD: Gradient-Coupled Anomaly Detection for Memory Poisoning in Retrieval-Augmented Agents","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2412.05579","citing_title":"LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methods","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01591","citing_title":"Led to Mislead: Adversarial Content Injection for Attacks on Neural Ranking Models","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18756","citing_title":"Towards Understanding the Robustness of Sparse Autoencoders","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03482","citing_title":"MEMSAD: Gradient-Coupled Anomaly Detection for Memory Poisoning in Retrieval-Augmented Agents","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VBYQDTRMQU7BSLKVWGH2JITEMK","json":"https://pith.science/pith/VBYQDTRMQU7BSLKVWGH2JITEMK.json","graph_json":"https://pith.science/api/pith-number/VBYQDTRMQU7BSLKVWGH2JITEMK/graph.json","events_json":"https://pith.science/api/pith-number/VBYQDTRMQU7BSLKVWGH2JITEMK/events.json","paper":"https://pith.science/paper/VBYQDTRM"},"agent_actions":{"view_html":"https://pith.science/pith/VBYQDTRMQU7BSLKVWGH2JITEMK","download_json":"https://pith.science/pith/VBYQDTRMQU7BSLKVWGH2JITEMK.json","view_paper":"https://pith.science/paper/VBYQDTRM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1712.06751&json=true","fetch_graph":"https://pith.science/api/pith-number/VBYQDTRMQU7BSLKVWGH2JITEMK/graph.json","fetch_events":"https://pith.science/api/pith-number/VBYQDTRMQU7BSLKVWGH2JITEMK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VBYQDTRMQU7BSLKVWGH2JITEMK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VBYQDTRMQU7BSLKVWGH2JITEMK/action/storage_attestation","attest_author":"https://pith.science/pith/VBYQDTRMQU7BSLKVWGH2JITEMK/action/author_attestation","sign_citation":"https://pith.science/pith/VBYQDTRMQU7BSLKVWGH2JITEMK/action/citation_signature","submit_replication":"https://pith.science/pith/VBYQDTRMQU7BSLKVWGH2JITEMK/action/replication_record"}},"created_at":"2026-05-18T00:15:05.143872+00:00","updated_at":"2026-05-18T00:15:05.143872+00:00"}