{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:5B7UNC4V2KMKK5JOZAI5G6F7JC","short_pith_number":"pith:5B7UNC4V","schema_version":"1.0","canonical_sha256":"e87f468b95d298a5752ec811d378bf48a0111d172d169c12c59ba2d1e60c0ec6","source":{"kind":"arxiv","id":"2311.08596","version":2},"attestation_state":"computed","paper":{"title":"Are You Sure? Challenging LLMs Leads to Performance Drops in The FlipFlop Experiment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Caiming Xiong, Chien-Sheng Wu, Lidiya Murakhovs'ka, Philippe Laban","submitted_at":"2023-11-14T23:40:22Z","abstract_excerpt":"The interactive nature of Large Language Models (LLMs) theoretically allows models to refine and improve their answers, yet systematic analysis of the multi-turn behavior of LLMs remains limited. In this paper, we propose the FlipFlop experiment: in the first round of the conversation, an LLM completes a classification task. In a second round, the LLM is challenged with a follow-up phrase like \"Are you sure?\", offering an opportunity for the model to reflect on its initial answer, and decide whether to confirm or flip its answer. A systematic study of ten LLMs on seven classification tasks rev"},"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":"2311.08596","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2023-11-14T23:40:22Z","cross_cats_sorted":[],"title_canon_sha256":"645a6771b3463ddda0b6bc36022dd517a8106fd54dfcaddec22e85c54039de4b","abstract_canon_sha256":"b3fda966c22589597f346230b7a40704e934f969018bf208fcc5ef47b12c607f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T07:47:45.688114Z","signature_b64":"CxLEaSqsPMftsRy36EhOBGrum1qjRV261zKaiwV19mNQQI61qSN4nyBc1/bCu035CW3o5BDoAq9chYmUgLOkCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e87f468b95d298a5752ec811d378bf48a0111d172d169c12c59ba2d1e60c0ec6","last_reissued_at":"2026-07-05T07:47:45.687547Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T07:47:45.687547Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Are You Sure? Challenging LLMs Leads to Performance Drops in The FlipFlop Experiment","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Caiming Xiong, Chien-Sheng Wu, Lidiya Murakhovs'ka, Philippe Laban","submitted_at":"2023-11-14T23:40:22Z","abstract_excerpt":"The interactive nature of Large Language Models (LLMs) theoretically allows models to refine and improve their answers, yet systematic analysis of the multi-turn behavior of LLMs remains limited. In this paper, we propose the FlipFlop experiment: in the first round of the conversation, an LLM completes a classification task. In a second round, the LLM is challenged with a follow-up phrase like \"Are you sure?\", offering an opportunity for the model to reflect on its initial answer, and decide whether to confirm or flip its answer. A systematic study of ten LLMs on seven classification tasks rev"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2311.08596","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/2311.08596/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":"2311.08596","created_at":"2026-07-05T07:47:45.687608+00:00"},{"alias_kind":"arxiv_version","alias_value":"2311.08596v2","created_at":"2026-07-05T07:47:45.687608+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2311.08596","created_at":"2026-07-05T07:47:45.687608+00:00"},{"alias_kind":"pith_short_12","alias_value":"5B7UNC4V2KMK","created_at":"2026-07-05T07:47:45.687608+00:00"},{"alias_kind":"pith_short_16","alias_value":"5B7UNC4V2KMKK5JO","created_at":"2026-07-05T07:47:45.687608+00:00"},{"alias_kind":"pith_short_8","alias_value":"5B7UNC4V","created_at":"2026-07-05T07:47:45.687608+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":10,"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":21,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18060","citing_title":"PseudoBench: Measuring How Agentic Auto-Research Fuels Pseudoscience","ref_index":125,"is_internal_anchor":false},{"citing_arxiv_id":"2605.21778","citing_title":"What Counts as AI Sycophancy? A Taxonomy and Expert Survey of a Fragmented Construct","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2508.16846","citing_title":"BASIL: Bayesian Assessment of Sycophancy in LLMs","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2601.10467","citing_title":"User Detection and Response Patterns of Sycophantic Behavior in Conversational AI","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2505.06120","citing_title":"LLMs Get Lost In Multi-Turn Conversation","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02668","citing_title":"Too Polite to Disagree: Understanding Sycophancy Propagation in Multi-Agent Systems","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09314","citing_title":"How LLMs Are Persuaded: A Few Attention Heads, Rerouted","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06476","citing_title":"Towards Emotion Consistency Analysis of Large Language Models in Emotional Conversational Contexts","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01048","citing_title":"Compared to What? Baselines and Metrics for Counterfactual Prompting","ref_index":55,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5B7UNC4V2KMKK5JOZAI5G6F7JC","json":"https://pith.science/pith/5B7UNC4V2KMKK5JOZAI5G6F7JC.json","graph_json":"https://pith.science/api/pith-number/5B7UNC4V2KMKK5JOZAI5G6F7JC/graph.json","events_json":"https://pith.science/api/pith-number/5B7UNC4V2KMKK5JOZAI5G6F7JC/events.json","paper":"https://pith.science/paper/5B7UNC4V"},"agent_actions":{"view_html":"https://pith.science/pith/5B7UNC4V2KMKK5JOZAI5G6F7JC","download_json":"https://pith.science/pith/5B7UNC4V2KMKK5JOZAI5G6F7JC.json","view_paper":"https://pith.science/paper/5B7UNC4V","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2311.08596&json=true","fetch_graph":"https://pith.science/api/pith-number/5B7UNC4V2KMKK5JOZAI5G6F7JC/graph.json","fetch_events":"https://pith.science/api/pith-number/5B7UNC4V2KMKK5JOZAI5G6F7JC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5B7UNC4V2KMKK5JOZAI5G6F7JC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5B7UNC4V2KMKK5JOZAI5G6F7JC/action/storage_attestation","attest_author":"https://pith.science/pith/5B7UNC4V2KMKK5JOZAI5G6F7JC/action/author_attestation","sign_citation":"https://pith.science/pith/5B7UNC4V2KMKK5JOZAI5G6F7JC/action/citation_signature","submit_replication":"https://pith.science/pith/5B7UNC4V2KMKK5JOZAI5G6F7JC/action/replication_record"}},"created_at":"2026-07-05T07:47:45.687608+00:00","updated_at":"2026-07-05T07:47:45.687608+00:00"}