{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:3OVQ2TMKJE7PXNOS6DVJGZQ2QB","short_pith_number":"pith:3OVQ2TMK","canonical_record":{"source":{"id":"2606.07237","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-05T13:07:11Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"fe8f710d52cfa2bbf33cb38e8520823bdbdf5fe0aa654837f66d9b9a5e030e06","abstract_canon_sha256":"1c785a1cad81bac43927030ceca8bdaaca9189992eec7611417357f495137d4e"},"schema_version":"1.0"},"canonical_sha256":"dbab0d4d8a493efbb5d2f0ea93661a8058b705ee736d86b720202cab39498930","source":{"kind":"arxiv","id":"2606.07237","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.07237","created_at":"2026-06-08T01:05:14Z"},{"alias_kind":"arxiv_version","alias_value":"2606.07237v1","created_at":"2026-06-08T01:05:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.07237","created_at":"2026-06-08T01:05:14Z"},{"alias_kind":"pith_short_12","alias_value":"3OVQ2TMKJE7P","created_at":"2026-06-08T01:05:14Z"},{"alias_kind":"pith_short_16","alias_value":"3OVQ2TMKJE7PXNOS","created_at":"2026-06-08T01:05:14Z"},{"alias_kind":"pith_short_8","alias_value":"3OVQ2TMK","created_at":"2026-06-08T01:05:14Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:3OVQ2TMKJE7PXNOS6DVJGZQ2QB","target":"record","payload":{"canonical_record":{"source":{"id":"2606.07237","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-05T13:07:11Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"fe8f710d52cfa2bbf33cb38e8520823bdbdf5fe0aa654837f66d9b9a5e030e06","abstract_canon_sha256":"1c785a1cad81bac43927030ceca8bdaaca9189992eec7611417357f495137d4e"},"schema_version":"1.0"},"canonical_sha256":"dbab0d4d8a493efbb5d2f0ea93661a8058b705ee736d86b720202cab39498930","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-08T01:05:14.749639Z","signature_b64":"q3MbW34m15EOinUaTp5sqpBe32sYQ77GGgzAbH8VC76vJqRlaTVdjBDz6tspur2ET7+EVKS4/73+Y25nVxUkCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dbab0d4d8a493efbb5d2f0ea93661a8058b705ee736d86b720202cab39498930","last_reissued_at":"2026-06-08T01:05:14.748742Z","signature_status":"signed_v1","first_computed_at":"2026-06-08T01:05:14.748742Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2606.07237","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-08T01:05:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"PlW1FEb3Yh3z08CvIgU35tC4ravU3vDmfNgSbG3FoT/Vo0GVE8/dLjRR4y3nV/ZuOzqWX3Rw6YyALgdxQJjPBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T07:51:15.825515Z"},"content_sha256":"f16a8cd181e6e23ab7c33191d90e7609c3757eed9951f328a1e9d1064bf34794","schema_version":"1.0","event_id":"sha256:f16a8cd181e6e23ab7c33191d90e7609c3757eed9951f328a1e9d1064bf34794"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:3OVQ2TMKJE7PXNOS6DVJGZQ2QB","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"When Large Language Models Fail in Healthcare: Evaluating Sensitivity to Prompt Variations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Mahdi Alkaeed","submitted_at":"2026-06-05T13:07:11Z","abstract_excerpt":"Large Language Models (LLMs) are increasingly used in healthcare for tasks such as clinical question answering, diagnosis support, and report summarization. Despite their promise, these models remain highly sensitive to subtle prompt perturbations, both lexical and syntactic, posing serious risks in safety-critical clinical applications. In this study, we conduct a systematic sensitivity analysis to evaluate the robustness of both general-purpose (e.g., GPT-3.5, Llama3) and medical-specific LLMs (e.g., ClinicalBERT, BioLlama3, BioBERT) using the MedMCQA benchmark. We categorize perturbations i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.07237","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/2606.07237/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-08T01:05:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5AVKN4LVaR1u2sTjvGI7lcAmzHCH39An0adlf9Tul0bmmKLx0Z6TzSo/YSf3QHUBC9TJBwxNt+Dq0nJOUgqwBw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T07:51:15.825903Z"},"content_sha256":"8b6201888dedc4be2eadccb7a63aa680ed014ac437d4e7e502e1c6b05fc3adcd","schema_version":"1.0","event_id":"sha256:8b6201888dedc4be2eadccb7a63aa680ed014ac437d4e7e502e1c6b05fc3adcd"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/3OVQ2TMKJE7PXNOS6DVJGZQ2QB/bundle.json","state_url":"https://pith.science/pith/3OVQ2TMKJE7PXNOS6DVJGZQ2QB/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/3OVQ2TMKJE7PXNOS6DVJGZQ2QB/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-03T07:51:15Z","links":{"resolver":"https://pith.science/pith/3OVQ2TMKJE7PXNOS6DVJGZQ2QB","bundle":"https://pith.science/pith/3OVQ2TMKJE7PXNOS6DVJGZQ2QB/bundle.json","state":"https://pith.science/pith/3OVQ2TMKJE7PXNOS6DVJGZQ2QB/state.json","well_known_bundle":"https://pith.science/.well-known/pith/3OVQ2TMKJE7PXNOS6DVJGZQ2QB/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:3OVQ2TMKJE7PXNOS6DVJGZQ2QB","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"1c785a1cad81bac43927030ceca8bdaaca9189992eec7611417357f495137d4e","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-05T13:07:11Z","title_canon_sha256":"fe8f710d52cfa2bbf33cb38e8520823bdbdf5fe0aa654837f66d9b9a5e030e06"},"schema_version":"1.0","source":{"id":"2606.07237","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.07237","created_at":"2026-06-08T01:05:14Z"},{"alias_kind":"arxiv_version","alias_value":"2606.07237v1","created_at":"2026-06-08T01:05:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.07237","created_at":"2026-06-08T01:05:14Z"},{"alias_kind":"pith_short_12","alias_value":"3OVQ2TMKJE7P","created_at":"2026-06-08T01:05:14Z"},{"alias_kind":"pith_short_16","alias_value":"3OVQ2TMKJE7PXNOS","created_at":"2026-06-08T01:05:14Z"},{"alias_kind":"pith_short_8","alias_value":"3OVQ2TMK","created_at":"2026-06-08T01:05:14Z"}],"graph_snapshots":[{"event_id":"sha256:8b6201888dedc4be2eadccb7a63aa680ed014ac437d4e7e502e1c6b05fc3adcd","target":"graph","created_at":"2026-06-08T01:05:14Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2606.07237/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Large Language Models (LLMs) are increasingly used in healthcare for tasks such as clinical question answering, diagnosis support, and report summarization. Despite their promise, these models remain highly sensitive to subtle prompt perturbations, both lexical and syntactic, posing serious risks in safety-critical clinical applications. In this study, we conduct a systematic sensitivity analysis to evaluate the robustness of both general-purpose (e.g., GPT-3.5, Llama3) and medical-specific LLMs (e.g., ClinicalBERT, BioLlama3, BioBERT) using the MedMCQA benchmark. We categorize perturbations i","authors_text":"Mahdi Alkaeed","cross_cats":["cs.AI","cs.LG"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-05T13:07:11Z","title":"When Large Language Models Fail in Healthcare: Evaluating Sensitivity to Prompt Variations"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.07237","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:f16a8cd181e6e23ab7c33191d90e7609c3757eed9951f328a1e9d1064bf34794","target":"record","created_at":"2026-06-08T01:05:14Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"1c785a1cad81bac43927030ceca8bdaaca9189992eec7611417357f495137d4e","cross_cats_sorted":["cs.AI","cs.LG"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-06-05T13:07:11Z","title_canon_sha256":"fe8f710d52cfa2bbf33cb38e8520823bdbdf5fe0aa654837f66d9b9a5e030e06"},"schema_version":"1.0","source":{"id":"2606.07237","kind":"arxiv","version":1}},"canonical_sha256":"dbab0d4d8a493efbb5d2f0ea93661a8058b705ee736d86b720202cab39498930","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"dbab0d4d8a493efbb5d2f0ea93661a8058b705ee736d86b720202cab39498930","first_computed_at":"2026-06-08T01:05:14.748742Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-08T01:05:14.748742Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"q3MbW34m15EOinUaTp5sqpBe32sYQ77GGgzAbH8VC76vJqRlaTVdjBDz6tspur2ET7+EVKS4/73+Y25nVxUkCw==","signature_status":"signed_v1","signed_at":"2026-06-08T01:05:14.749639Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.07237","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f16a8cd181e6e23ab7c33191d90e7609c3757eed9951f328a1e9d1064bf34794","sha256:8b6201888dedc4be2eadccb7a63aa680ed014ac437d4e7e502e1c6b05fc3adcd"],"state_sha256":"b927e8e532c885325d4c03bf9da10b61e31a503d105ef2ef0c3f6ee2afa27c04"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kH8DCm3ANadmu4qpfCcmEYtvIff2PIYQcaP0tI/6mUSlTWzvFliyWaXQ2T7Ux2xRgA8hxxBGGcsweDlZY07ZAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-03T07:51:15.829113Z","bundle_sha256":"bd85479cff53ab38e9c4142344cbb69bf69ffef8c25d3cba24a56c9cb11a7855"}}