{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:SHDFL7Z3LQQTETGDDNWM54IUCI","short_pith_number":"pith:SHDFL7Z3","schema_version":"1.0","canonical_sha256":"91c655ff3b5c21324cc31b6ccef11412224257683572f962f1f5b4e8a42e4313","source":{"kind":"arxiv","id":"2305.14950","version":2},"attestation_state":"computed","paper":{"title":"Adversarial Demonstration Attacks on Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chaowei Xiao, Jiongxiao Wang, Keun Hee Park, Muhao Chen, Zhaoheng Zheng, Zhuofeng Wu, Zhuojun Jiang, Zichen Liu","submitted_at":"2023-05-24T09:40:56Z","abstract_excerpt":"With the emergence of more powerful large language models (LLMs), such as ChatGPT and GPT-4, in-context learning (ICL) has gained significant prominence in leveraging these models for specific tasks by utilizing data-label pairs as precondition prompts. While incorporating demonstrations can greatly enhance the performance of LLMs across various tasks, it may introduce a new security concern: attackers can manipulate only the demonstrations without changing the input to perform an attack. In this paper, we investigate the security concern of ICL from an adversarial perspective, focusing on 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":"2305.14950","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-05-24T09:40:56Z","cross_cats_sorted":["cs.AI","cs.CR","cs.LG"],"title_canon_sha256":"9746d8d9f7353e8d416bba1938b60c83f2c6994fe6926c2dcbc3df8b7101f654","abstract_canon_sha256":"1b2f6c012af125557059a9586fd995928a327a91526fb4a06cc2768df89e01ef"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:00:44.694604Z","signature_b64":"xdmZr/LEPlYeytoBNHh3nCHtIK6HjixePJgy3gFyjg9qVZQ2ztqOOCRuv7GdGVX4h/ZjMXYMvQbFElwCX4vQDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"91c655ff3b5c21324cc31b6ccef11412224257683572f962f1f5b4e8a42e4313","last_reissued_at":"2026-07-05T07:00:44.694113Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:00:44.694113Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Adversarial Demonstration Attacks on Large Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CR","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chaowei Xiao, Jiongxiao Wang, Keun Hee Park, Muhao Chen, Zhaoheng Zheng, Zhuofeng Wu, Zhuojun Jiang, Zichen Liu","submitted_at":"2023-05-24T09:40:56Z","abstract_excerpt":"With the emergence of more powerful large language models (LLMs), such as ChatGPT and GPT-4, in-context learning (ICL) has gained significant prominence in leveraging these models for specific tasks by utilizing data-label pairs as precondition prompts. While incorporating demonstrations can greatly enhance the performance of LLMs across various tasks, it may introduce a new security concern: attackers can manipulate only the demonstrations without changing the input to perform an attack. In this paper, we investigate the security concern of ICL from an adversarial perspective, focusing on the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.14950","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/2305.14950/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":"2305.14950","created_at":"2026-07-05T07:00:44.694170+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.14950v2","created_at":"2026-07-05T07:00:44.694170+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.14950","created_at":"2026-07-05T07:00:44.694170+00:00"},{"alias_kind":"pith_short_12","alias_value":"SHDFL7Z3LQQT","created_at":"2026-07-05T07:00:44.694170+00:00"},{"alias_kind":"pith_short_16","alias_value":"SHDFL7Z3LQQTETGD","created_at":"2026-07-05T07:00:44.694170+00:00"},{"alias_kind":"pith_short_8","alias_value":"SHDFL7Z3","created_at":"2026-07-05T07:00:44.694170+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.26350","citing_title":"When Correct Demonstrations Hurt: Rethinking the Role of Exemplars in In-Context Learning","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2502.05206","citing_title":"Safety at Scale: A Comprehensive Survey of Large Model and Agent Safety","ref_index":65,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15865","citing_title":"From Text to DSL: Evaluating Grammar-Based Model Generation Using Open LLMs","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2510.16558","citing_title":"A First Look at the Security Issues in the Model Context Protocol Ecosystem","ref_index":48,"is_internal_anchor":false},{"citing_arxiv_id":"2407.04295","citing_title":"Jailbreak Attacks and Defenses Against Large Language Models: A Survey","ref_index":95,"is_internal_anchor":false},{"citing_arxiv_id":"2310.03684","citing_title":"SmoothLLM: Defending Large Language Models Against Jailbreaking Attacks","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2502.18864","citing_title":"Towards an AI co-scientist","ref_index":17,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SHDFL7Z3LQQTETGDDNWM54IUCI","json":"https://pith.science/pith/SHDFL7Z3LQQTETGDDNWM54IUCI.json","graph_json":"https://pith.science/api/pith-number/SHDFL7Z3LQQTETGDDNWM54IUCI/graph.json","events_json":"https://pith.science/api/pith-number/SHDFL7Z3LQQTETGDDNWM54IUCI/events.json","paper":"https://pith.science/paper/SHDFL7Z3"},"agent_actions":{"view_html":"https://pith.science/pith/SHDFL7Z3LQQTETGDDNWM54IUCI","download_json":"https://pith.science/pith/SHDFL7Z3LQQTETGDDNWM54IUCI.json","view_paper":"https://pith.science/paper/SHDFL7Z3","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.14950&json=true","fetch_graph":"https://pith.science/api/pith-number/SHDFL7Z3LQQTETGDDNWM54IUCI/graph.json","fetch_events":"https://pith.science/api/pith-number/SHDFL7Z3LQQTETGDDNWM54IUCI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SHDFL7Z3LQQTETGDDNWM54IUCI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SHDFL7Z3LQQTETGDDNWM54IUCI/action/storage_attestation","attest_author":"https://pith.science/pith/SHDFL7Z3LQQTETGDDNWM54IUCI/action/author_attestation","sign_citation":"https://pith.science/pith/SHDFL7Z3LQQTETGDDNWM54IUCI/action/citation_signature","submit_replication":"https://pith.science/pith/SHDFL7Z3LQQTETGDDNWM54IUCI/action/replication_record"}},"created_at":"2026-07-05T07:00:44.694170+00:00","updated_at":"2026-07-05T07:00:44.694170+00:00"}