{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:BB2BVX65WSAT3O5ZL557GA7JJC","short_pith_number":"pith:BB2BVX65","schema_version":"1.0","canonical_sha256":"08741adfddb4813dbbb95f7bf303e9488afd924b1c2f8d086b852718e103b3d0","source":{"kind":"arxiv","id":"2312.04004","version":2},"attestation_state":"computed","paper":{"title":"Occlusion-based Detection of Trojan-triggering Inputs in Large Language Models of Code","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Aftab Hussain, Bowen Xu, Md Rafiqul Islam Rabin, Mohammad Amin Alipour, Toufique Ahmed","submitted_at":"2023-12-07T02:44:35Z","abstract_excerpt":"Large language models (LLMs) are becoming an integrated part of software development. These models are trained on large datasets for code, where it is hard to verify each data point. Therefore, a potential attack surface can be to inject poisonous data into the training data to make models vulnerable, aka trojaned. It can pose a significant threat by hiding manipulative behaviors inside models, leading to compromising the integrity of the models in downstream tasks.\n  In this paper, we propose an occlusion-based human-in-the-loop technique, OSeql, to distinguish trojan-triggering inputs of cod"},"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":"2312.04004","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.SE","submitted_at":"2023-12-07T02:44:35Z","cross_cats_sorted":[],"title_canon_sha256":"10e324c7e988844e432292370eab1917b1eb5b475c157a3a9087fd077cb34979","abstract_canon_sha256":"3a9b8593aab1caa6565fc26f1f632f0fe0970b4b6733bba7391c06ccb1d35895"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:22:27.529463Z","signature_b64":"7NFUuZhGGEVxFPCv9KC4yYoAVPSlDzqOLOUFupZGfyokeNjVsYMpUVvdM5O1U/CAjg5DOpHh8vkJqEBMvypkBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"08741adfddb4813dbbb95f7bf303e9488afd924b1c2f8d086b852718e103b3d0","last_reissued_at":"2026-07-05T07:22:27.528926Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:22:27.528926Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Occlusion-based Detection of Trojan-triggering Inputs in Large Language Models of Code","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.SE","authors_text":"Aftab Hussain, Bowen Xu, Md Rafiqul Islam Rabin, Mohammad Amin Alipour, Toufique Ahmed","submitted_at":"2023-12-07T02:44:35Z","abstract_excerpt":"Large language models (LLMs) are becoming an integrated part of software development. These models are trained on large datasets for code, where it is hard to verify each data point. Therefore, a potential attack surface can be to inject poisonous data into the training data to make models vulnerable, aka trojaned. It can pose a significant threat by hiding manipulative behaviors inside models, leading to compromising the integrity of the models in downstream tasks.\n  In this paper, we propose an occlusion-based human-in-the-loop technique, OSeql, to distinguish trojan-triggering inputs of cod"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2312.04004","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/2312.04004/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":"2312.04004","created_at":"2026-07-05T07:22:27.528999+00:00"},{"alias_kind":"arxiv_version","alias_value":"2312.04004v2","created_at":"2026-07-05T07:22:27.528999+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2312.04004","created_at":"2026-07-05T07:22:27.528999+00:00"},{"alias_kind":"pith_short_12","alias_value":"BB2BVX65WSAT","created_at":"2026-07-05T07:22:27.528999+00:00"},{"alias_kind":"pith_short_16","alias_value":"BB2BVX65WSAT3O5Z","created_at":"2026-07-05T07:22:27.528999+00:00"},{"alias_kind":"pith_short_8","alias_value":"BB2BVX65","created_at":"2026-07-05T07:22:27.528999+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.10846","citing_title":"Securing Code Understanding: Detecting Natural Backdoor Vulnerability in Code Language Models","ref_index":25,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BB2BVX65WSAT3O5ZL557GA7JJC","json":"https://pith.science/pith/BB2BVX65WSAT3O5ZL557GA7JJC.json","graph_json":"https://pith.science/api/pith-number/BB2BVX65WSAT3O5ZL557GA7JJC/graph.json","events_json":"https://pith.science/api/pith-number/BB2BVX65WSAT3O5ZL557GA7JJC/events.json","paper":"https://pith.science/paper/BB2BVX65"},"agent_actions":{"view_html":"https://pith.science/pith/BB2BVX65WSAT3O5ZL557GA7JJC","download_json":"https://pith.science/pith/BB2BVX65WSAT3O5ZL557GA7JJC.json","view_paper":"https://pith.science/paper/BB2BVX65","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2312.04004&json=true","fetch_graph":"https://pith.science/api/pith-number/BB2BVX65WSAT3O5ZL557GA7JJC/graph.json","fetch_events":"https://pith.science/api/pith-number/BB2BVX65WSAT3O5ZL557GA7JJC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BB2BVX65WSAT3O5ZL557GA7JJC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BB2BVX65WSAT3O5ZL557GA7JJC/action/storage_attestation","attest_author":"https://pith.science/pith/BB2BVX65WSAT3O5ZL557GA7JJC/action/author_attestation","sign_citation":"https://pith.science/pith/BB2BVX65WSAT3O5ZL557GA7JJC/action/citation_signature","submit_replication":"https://pith.science/pith/BB2BVX65WSAT3O5ZL557GA7JJC/action/replication_record"}},"created_at":"2026-07-05T07:22:27.528999+00:00","updated_at":"2026-07-05T07:22:27.528999+00:00"}