{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:PJLR7YHMCEGYLTOBTPWLZXJPIX","short_pith_number":"pith:PJLR7YHM","schema_version":"1.0","canonical_sha256":"7a571fe0ec110d85cdc19becbcdd2f45d2b883abc82ebbbaecbc6fa9f617ac21","source":{"kind":"arxiv","id":"2509.03986","version":1},"attestation_state":"computed","paper":{"title":"Promptception: How Sensitive Are Large Multimodal Models to Prompts?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Mohamed Insaf Ismithdeen, Muhammad Uzair Khattak, Salman Khan","submitted_at":"2025-09-04T08:13:06Z","abstract_excerpt":"Despite the success of Large Multimodal Models (LMMs) in recent years, prompt design for LMMs in Multiple-Choice Question Answering (MCQA) remains poorly understood. We show that even minor variations in prompt phrasing and structure can lead to accuracy deviations of up to 15% for certain prompts and models. This variability poses a challenge for transparent and fair LMM evaluation, as models often report their best-case performance using carefully selected prompts. To address this, we introduce Promptception, a systematic framework for evaluating prompt sensitivity in LMMs. It consists of 61"},"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":"2509.03986","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-09-04T08:13:06Z","cross_cats_sorted":["cs.AI","cs.CL","cs.LG"],"title_canon_sha256":"eadac637eb68da13373ae92e4ef824fb8244c10123b8b52d21ad086c275dbe75","abstract_canon_sha256":"fddd75dbb09313db81bd9bba2a6f2ecf1c48b80a939c48c8962114d256f7ce06"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:04:51.200121Z","signature_b64":"+8dlO0M7B7J2m4HMCWC1UvvdtroqOXHQNfAPlbWHWbMj2q4IKoNfW66i+YS3Ni3h+6gD3DH0p4CDpLYohylHDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7a571fe0ec110d85cdc19becbcdd2f45d2b883abc82ebbbaecbc6fa9f617ac21","last_reissued_at":"2026-07-05T12:04:51.199610Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:04:51.199610Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Promptception: How Sensitive Are Large Multimodal Models to Prompts?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.LG"],"primary_cat":"cs.CV","authors_text":"Mohamed Insaf Ismithdeen, Muhammad Uzair Khattak, Salman Khan","submitted_at":"2025-09-04T08:13:06Z","abstract_excerpt":"Despite the success of Large Multimodal Models (LMMs) in recent years, prompt design for LMMs in Multiple-Choice Question Answering (MCQA) remains poorly understood. We show that even minor variations in prompt phrasing and structure can lead to accuracy deviations of up to 15% for certain prompts and models. This variability poses a challenge for transparent and fair LMM evaluation, as models often report their best-case performance using carefully selected prompts. To address this, we introduce Promptception, a systematic framework for evaluating prompt sensitivity in LMMs. It consists of 61"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2509.03986","kind":"arxiv","version":1},"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/2509.03986/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":"2509.03986","created_at":"2026-07-05T12:04:51.199677+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.03986v1","created_at":"2026-07-05T12:04:51.199677+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.03986","created_at":"2026-07-05T12:04:51.199677+00:00"},{"alias_kind":"pith_short_12","alias_value":"PJLR7YHMCEGY","created_at":"2026-07-05T12:04:51.199677+00:00"},{"alias_kind":"pith_short_16","alias_value":"PJLR7YHMCEGYLTOB","created_at":"2026-07-05T12:04:51.199677+00:00"},{"alias_kind":"pith_short_8","alias_value":"PJLR7YHM","created_at":"2026-07-05T12:04:51.199677+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26079","citing_title":"Same Evidence, Different Answer: Auditing Order Sensitivity in Multimodal Large Language Models","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30646","citing_title":"Same Patient, Different Words, Different Diagnosis? Evaluating Semantic Stability in Clinical LLMs","ref_index":37,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PJLR7YHMCEGYLTOBTPWLZXJPIX","json":"https://pith.science/pith/PJLR7YHMCEGYLTOBTPWLZXJPIX.json","graph_json":"https://pith.science/api/pith-number/PJLR7YHMCEGYLTOBTPWLZXJPIX/graph.json","events_json":"https://pith.science/api/pith-number/PJLR7YHMCEGYLTOBTPWLZXJPIX/events.json","paper":"https://pith.science/paper/PJLR7YHM"},"agent_actions":{"view_html":"https://pith.science/pith/PJLR7YHMCEGYLTOBTPWLZXJPIX","download_json":"https://pith.science/pith/PJLR7YHMCEGYLTOBTPWLZXJPIX.json","view_paper":"https://pith.science/paper/PJLR7YHM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.03986&json=true","fetch_graph":"https://pith.science/api/pith-number/PJLR7YHMCEGYLTOBTPWLZXJPIX/graph.json","fetch_events":"https://pith.science/api/pith-number/PJLR7YHMCEGYLTOBTPWLZXJPIX/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PJLR7YHMCEGYLTOBTPWLZXJPIX/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PJLR7YHMCEGYLTOBTPWLZXJPIX/action/storage_attestation","attest_author":"https://pith.science/pith/PJLR7YHMCEGYLTOBTPWLZXJPIX/action/author_attestation","sign_citation":"https://pith.science/pith/PJLR7YHMCEGYLTOBTPWLZXJPIX/action/citation_signature","submit_replication":"https://pith.science/pith/PJLR7YHMCEGYLTOBTPWLZXJPIX/action/replication_record"}},"created_at":"2026-07-05T12:04:51.199677+00:00","updated_at":"2026-07-05T12:04:51.199677+00:00"}