{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:XSPRTKSIRXEWQ7TCQYHVWLZC4O","short_pith_number":"pith:XSPRTKSI","schema_version":"1.0","canonical_sha256":"bc9f19aa488dc9687e62860f5b2f22e380fd78fe815e3774e48fa8e09dbc04d8","source":{"kind":"arxiv","id":"2402.12483","version":2},"attestation_state":"computed","paper":{"title":"Artifacts or Abduction: How Do LLMs Answer Multiple-Choice Questions Without the Question?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Abhilasha Ravichander, Nishant Balepur, Rachel Rudinger","submitted_at":"2024-02-19T19:38:58Z","abstract_excerpt":"Multiple-choice question answering (MCQA) is often used to evaluate large language models (LLMs). To see if MCQA assesses LLMs as intended, we probe if LLMs can perform MCQA with choices-only prompts, where models must select the correct answer only from the choices. In three MCQA datasets and four LLMs, this prompt bests a majority baseline in 11/12 cases, with up to 0.33 accuracy gain. To help explain this behavior, we conduct an in-depth, black-box analysis on memorization, choice dynamics, and question inference. Our key findings are threefold. First, we find no evidence that the choices-o"},"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":"2402.12483","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2024-02-19T19:38:58Z","cross_cats_sorted":[],"title_canon_sha256":"a94957d4f3559c4986bfd8121d4c50b1de9ce3c2856934c5059ebf827e683d46","abstract_canon_sha256":"89d60ff9b4d7fa3b6240eb08a92b846ecbf3bcc383c71b0287c01803599a5480"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:28:59.760514Z","signature_b64":"txNI/f20AyVpi+nVpc3iA5zK0p5EoVgs68B+Y0yJWqIw/N2WJ5PPbKxAIF4V7P01Wkq1qvvXYxlRx8+PRGtPAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bc9f19aa488dc9687e62860f5b2f22e380fd78fe815e3774e48fa8e09dbc04d8","last_reissued_at":"2026-07-05T08:28:59.760025Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:28:59.760025Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Artifacts or Abduction: How Do LLMs Answer Multiple-Choice Questions Without the Question?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Abhilasha Ravichander, Nishant Balepur, Rachel Rudinger","submitted_at":"2024-02-19T19:38:58Z","abstract_excerpt":"Multiple-choice question answering (MCQA) is often used to evaluate large language models (LLMs). To see if MCQA assesses LLMs as intended, we probe if LLMs can perform MCQA with choices-only prompts, where models must select the correct answer only from the choices. In three MCQA datasets and four LLMs, this prompt bests a majority baseline in 11/12 cases, with up to 0.33 accuracy gain. To help explain this behavior, we conduct an in-depth, black-box analysis on memorization, choice dynamics, and question inference. Our key findings are threefold. First, we find no evidence that the choices-o"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.12483","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/2402.12483/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":"2402.12483","created_at":"2026-07-05T08:28:59.760086+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.12483v2","created_at":"2026-07-05T08:28:59.760086+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.12483","created_at":"2026-07-05T08:28:59.760086+00:00"},{"alias_kind":"pith_short_12","alias_value":"XSPRTKSIRXEW","created_at":"2026-07-05T08:28:59.760086+00:00"},{"alias_kind":"pith_short_16","alias_value":"XSPRTKSIRXEWQ7TC","created_at":"2026-07-05T08:28:59.760086+00:00"},{"alias_kind":"pith_short_8","alias_value":"XSPRTKSI","created_at":"2026-07-05T08:28:59.760086+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.12813","citing_title":"REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations","ref_index":65,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XSPRTKSIRXEWQ7TCQYHVWLZC4O","json":"https://pith.science/pith/XSPRTKSIRXEWQ7TCQYHVWLZC4O.json","graph_json":"https://pith.science/api/pith-number/XSPRTKSIRXEWQ7TCQYHVWLZC4O/graph.json","events_json":"https://pith.science/api/pith-number/XSPRTKSIRXEWQ7TCQYHVWLZC4O/events.json","paper":"https://pith.science/paper/XSPRTKSI"},"agent_actions":{"view_html":"https://pith.science/pith/XSPRTKSIRXEWQ7TCQYHVWLZC4O","download_json":"https://pith.science/pith/XSPRTKSIRXEWQ7TCQYHVWLZC4O.json","view_paper":"https://pith.science/paper/XSPRTKSI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.12483&json=true","fetch_graph":"https://pith.science/api/pith-number/XSPRTKSIRXEWQ7TCQYHVWLZC4O/graph.json","fetch_events":"https://pith.science/api/pith-number/XSPRTKSIRXEWQ7TCQYHVWLZC4O/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XSPRTKSIRXEWQ7TCQYHVWLZC4O/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XSPRTKSIRXEWQ7TCQYHVWLZC4O/action/storage_attestation","attest_author":"https://pith.science/pith/XSPRTKSIRXEWQ7TCQYHVWLZC4O/action/author_attestation","sign_citation":"https://pith.science/pith/XSPRTKSIRXEWQ7TCQYHVWLZC4O/action/citation_signature","submit_replication":"https://pith.science/pith/XSPRTKSIRXEWQ7TCQYHVWLZC4O/action/replication_record"}},"created_at":"2026-07-05T08:28:59.760086+00:00","updated_at":"2026-07-05T08:28:59.760086+00:00"}