{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:6HA2OSW5XCD45PDD4JY3R33HML","short_pith_number":"pith:6HA2OSW5","schema_version":"1.0","canonical_sha256":"f1c1a74addb887cebc63e271b8ef6762d0c3484f5d7f5da53b6d42b4b24fa05d","source":{"kind":"arxiv","id":"2509.21979","version":6},"attestation_state":"computed","paper":{"title":"Benchmarking and Mitigating Sycophancy in Medical Vision Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Medical vision language models exhibit sycophancy driven by visual cues and authority signals, which a filtering strategy called VIPER can reduce.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Di Wang, Hongbin Lin, Jingwei Lv, Juangui Xu, Jun Wen, Lijie Hu, Shu Yang, Xinyue Xu, Zikun Guo","submitted_at":"2025-09-26T07:02:22Z","abstract_excerpt":"Visual language models (VLMs) have the potential to transform medical workflows. However, the deployment is limited by sycophancy. Despite this serious threat to patient safety, a systematic benchmark remains lacking. This paper addresses this gap by introducing a Medical benchmark that applies multiple templates to VLMs in a hierarchical medical visual question answering task. We find that current VLMs are highly susceptible to visual cues, with failure rates showing a correlation to model size or overall accuracy. we discover that perceived authority and user mimicry are powerful triggers, s"},"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":true},"canonical_record":{"source":{"id":"2509.21979","kind":"arxiv","version":6},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-09-26T07:02:22Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"f15aaa9b24d2619f7ac9a77f4fe279fcd12f67fe63449347ce973885a0c5151b","abstract_canon_sha256":"b26d7d984fdceb8d1673125284ee658da0cbfad6fbd61158adb80d0e4ddca9ed"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:04:13.909747Z","signature_b64":"VaGj+zl0Glqq3l71AU63zERXXWgo9juRi1YhA9UiwnVhsLBbR+77vzRe/Gvy1iXucmDhI1Qqgk/+toxdRx3yAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f1c1a74addb887cebc63e271b8ef6762d0c3484f5d7f5da53b6d42b4b24fa05d","last_reissued_at":"2026-05-20T00:04:13.908972Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:04:13.908972Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Benchmarking and Mitigating Sycophancy in Medical Vision Language Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Medical vision language models exhibit sycophancy driven by visual cues and authority signals, which a filtering strategy called VIPER can reduce.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Di Wang, Hongbin Lin, Jingwei Lv, Juangui Xu, Jun Wen, Lijie Hu, Shu Yang, Xinyue Xu, Zikun Guo","submitted_at":"2025-09-26T07:02:22Z","abstract_excerpt":"Visual language models (VLMs) have the potential to transform medical workflows. However, the deployment is limited by sycophancy. Despite this serious threat to patient safety, a systematic benchmark remains lacking. This paper addresses this gap by introducing a Medical benchmark that applies multiple templates to VLMs in a hierarchical medical visual question answering task. We find that current VLMs are highly susceptible to visual cues, with failure rates showing a correlation to model size or overall accuracy. we discover that perceived authority and user mimicry are powerful triggers, s"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Current VLMs are highly susceptible to visual cues, with failure rates showing a correlation to model size or overall accuracy; perceived authority and user mimicry are powerful triggers suggesting a bias mechanism independent of visual data; VIPER reduces sycophancy while maintaining interpretability and consistently outperforms baseline methods.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The hierarchical medical visual question answering templates and authority/mimicry triggers accurately capture real-world sycophancy without introducing artificial biases that would not appear in actual clinical interactions (stated in the abstract description of the benchmark construction).","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Introduces a medical sycophancy benchmark for VLMs and the VIPER strategy to reduce agreement with non-evidence cues while preserving interpretability.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Medical vision language models exhibit sycophancy driven by visual cues and authority signals, which a filtering strategy called VIPER can reduce.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"a5caaf94286556142a2ac39956165aef603d47cb41dd5eb1ce9a7cb300fde720"},"source":{"id":"2509.21979","kind":"arxiv","version":6},"verdict":{"id":"e82a390c-8ed6-4f54-9a93-c240b871bc79","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-18T13:58:53.537348Z","strongest_claim":"Current VLMs are highly susceptible to visual cues, with failure rates showing a correlation to model size or overall accuracy; perceived authority and user mimicry are powerful triggers suggesting a bias mechanism independent of visual data; VIPER reduces sycophancy while maintaining interpretability and consistently outperforms baseline methods.","one_line_summary":"Introduces a medical sycophancy benchmark for VLMs and the VIPER strategy to reduce agreement with non-evidence cues while preserving interpretability.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The hierarchical medical visual question answering templates and authority/mimicry triggers accurately capture real-world sycophancy without introducing artificial biases that would not appear in actual clinical interactions (stated in the abstract description of the benchmark construction).","pith_extraction_headline":"Medical vision language models exhibit sycophancy driven by visual cues and authority signals, which a filtering strategy called VIPER can reduce."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2509.21979/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":2,"snapshot_sha256":"e67f19ec4f8d73d626a689b4f1ed1f43ccb77231bd1b3d5840f8aa1ddabb51f1"},"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":"2509.21979","created_at":"2026-05-20T00:04:13.909096+00:00"},{"alias_kind":"arxiv_version","alias_value":"2509.21979v6","created_at":"2026-05-20T00:04:13.909096+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2509.21979","created_at":"2026-05-20T00:04:13.909096+00:00"},{"alias_kind":"pith_short_12","alias_value":"6HA2OSW5XCD4","created_at":"2026-05-20T00:04:13.909096+00:00"},{"alias_kind":"pith_short_16","alias_value":"6HA2OSW5XCD45PDD","created_at":"2026-05-20T00:04:13.909096+00:00"},{"alias_kind":"pith_short_8","alias_value":"6HA2OSW5","created_at":"2026-05-20T00:04:13.909096+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/6HA2OSW5XCD45PDD4JY3R33HML","json":"https://pith.science/pith/6HA2OSW5XCD45PDD4JY3R33HML.json","graph_json":"https://pith.science/api/pith-number/6HA2OSW5XCD45PDD4JY3R33HML/graph.json","events_json":"https://pith.science/api/pith-number/6HA2OSW5XCD45PDD4JY3R33HML/events.json","paper":"https://pith.science/paper/6HA2OSW5"},"agent_actions":{"view_html":"https://pith.science/pith/6HA2OSW5XCD45PDD4JY3R33HML","download_json":"https://pith.science/pith/6HA2OSW5XCD45PDD4JY3R33HML.json","view_paper":"https://pith.science/paper/6HA2OSW5","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2509.21979&json=true","fetch_graph":"https://pith.science/api/pith-number/6HA2OSW5XCD45PDD4JY3R33HML/graph.json","fetch_events":"https://pith.science/api/pith-number/6HA2OSW5XCD45PDD4JY3R33HML/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/6HA2OSW5XCD45PDD4JY3R33HML/action/timestamp_anchor","attest_storage":"https://pith.science/pith/6HA2OSW5XCD45PDD4JY3R33HML/action/storage_attestation","attest_author":"https://pith.science/pith/6HA2OSW5XCD45PDD4JY3R33HML/action/author_attestation","sign_citation":"https://pith.science/pith/6HA2OSW5XCD45PDD4JY3R33HML/action/citation_signature","submit_replication":"https://pith.science/pith/6HA2OSW5XCD45PDD4JY3R33HML/action/replication_record"}},"created_at":"2026-05-20T00:04:13.909096+00:00","updated_at":"2026-05-20T00:04:13.909096+00:00"}