{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:NZ6UC77QVAQSZA2RNPTWACUPZ2","short_pith_number":"pith:NZ6UC77Q","canonical_record":{"source":{"id":"2608.04006","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-08-04T17:59:20Z","cross_cats_sorted":[],"title_canon_sha256":"cc737e574cbf679df8661269a705d49992ab92c65a1c278eaf0dadb67a2b4b7e","abstract_canon_sha256":"18d665843be2eb4d206004ddbbeb53fa9ebb83c88e5873fb623b9205c8bf98c7"},"schema_version":"1.0"},"canonical_sha256":"6e7d417ff0a8212c83516be7600a8fce8f2b3d8846d29bfb967b9890cd48c9b6","source":{"kind":"arxiv","id":"2608.04006","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.04006","created_at":"2026-08-05T01:37:41Z"},{"alias_kind":"arxiv_version","alias_value":"2608.04006v1","created_at":"2026-08-05T01:37:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.04006","created_at":"2026-08-05T01:37:41Z"},{"alias_kind":"pith_short_12","alias_value":"NZ6UC77QVAQS","created_at":"2026-08-05T01:37:41Z"},{"alias_kind":"pith_short_16","alias_value":"NZ6UC77QVAQSZA2R","created_at":"2026-08-05T01:37:41Z"},{"alias_kind":"pith_short_8","alias_value":"NZ6UC77Q","created_at":"2026-08-05T01:37:41Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:NZ6UC77QVAQSZA2RNPTWACUPZ2","target":"record","payload":{"canonical_record":{"source":{"id":"2608.04006","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-08-04T17:59:20Z","cross_cats_sorted":[],"title_canon_sha256":"cc737e574cbf679df8661269a705d49992ab92c65a1c278eaf0dadb67a2b4b7e","abstract_canon_sha256":"18d665843be2eb4d206004ddbbeb53fa9ebb83c88e5873fb623b9205c8bf98c7"},"schema_version":"1.0"},"canonical_sha256":"6e7d417ff0a8212c83516be7600a8fce8f2b3d8846d29bfb967b9890cd48c9b6","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-08-05T01:37:41.124630Z","signature_b64":"GGCt1263R0BTIy/xR5k/RTWNKm+ah1Dot9ebuYWsVPBzRciuQT0Jfv5sJWaLOjGt3/E0g/5bQFh3ggon/fKPCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6e7d417ff0a8212c83516be7600a8fce8f2b3d8846d29bfb967b9890cd48c9b6","last_reissued_at":"2026-08-05T01:37:41.123137Z","signature_status":"signed_v1","first_computed_at":"2026-08-05T01:37:41.123137Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2608.04006","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":"2026-08-05T01:37:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"cltj9aQam+2t1rr9NBcELxOP7GY6uBOZUIw0mTL+/6+oLeRxI4pB2apjaIO9mQf9YXte8ZITebfS51E2slnGBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T14:13:34.210163Z"},"content_sha256":"1b90c61caa227ca42a4bb51b2ae83798a6f7f5529dc2e14373c001eab4d1cfcb","schema_version":"1.0","event_id":"sha256:1b90c61caa227ca42a4bb51b2ae83798a6f7f5529dc2e14373c001eab4d1cfcb"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:NZ6UC77QVAQSZA2RNPTWACUPZ2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Calibrating Trustworthiness: Co-Designing Metrics and Visualizations for Evaluating LLMs in Education","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.HC","authors_text":"Adam Coscia, Alex Endert, Langdon Holmes, Scott Crossley, Sujata Duwal","submitted_at":"2026-08-04T17:59:20Z","abstract_excerpt":"LLMs are reshaping educational technology, yet evaluating their responses for pedagogical alignment remains underexplored, relying heavily on the expertise of learning engineers building the technology. To bridge this gap, we explore trustworthiness as a structured lens for evaluation, leveraging existing measures of LLM trustworthiness to systematically identify potential pedagogical disruptions. Through a longitudinal co-design process with learning engineers developing an LLM-powered digital textbook, we: (1) co-constructed five trustworthiness metrics comprising 20 measures tailored to ped"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.04006","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/2608.04006/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":"2026-08-05T01:37:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"WU7YC7kC+uw6zrgdIWnTa+X5dTX3SiWnV521Egp+Hq7Ic6F2x1lm7Lo/7Sf6y/AikwOhH6/IaRkWClRgq3S7Bw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T14:13:34.211131Z"},"content_sha256":"8859e1090884d9b5215c57061d584980e139e20facb7e000a3c8ac97b187ab46","schema_version":"1.0","event_id":"sha256:8859e1090884d9b5215c57061d584980e139e20facb7e000a3c8ac97b187ab46"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:NZ6UC77QVAQSZA2RNPTWACUPZ2","target":"integrity","payload":{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1109/TVCG.2012.213) was visible in the surrounding text but could not be confirmed against doi.org as printed.","snippet":"Michael Sedlmair, Miriah Meyer, and Tamara Munzner. 2012. Design Study Methodology: Reflections from the Trenches and the Stacks.IEEE Transactions on Visualization and Computer Graphics18, 12 (2012), 2431–2440. doi:10.1109/TVCG. 2012.213","arxiv_id":"2608.04006","detector":"doi_compliance","evidence":{"ref_index":52,"verdict_class":"incontrovertible","resolved_title":null,"printed_excerpt":"10.1109/tvcg","reconstructed_doi":"10.1109/TVCG.2012.213"},"severity":"advisory","ref_index":52,"audited_at":"2026-08-05T04:18:05.690783Z","event_type":"pith.integrity.v1","detected_doi":"10.1109/TVCG.2012.213","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"recoverable_identifier","evidence_hash":"c108f7415bda8a26885948b7c6f9a256fd02e352d25451ad8371e089ee099afd","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.1.0","detected_arxiv_id":null,"integrity_event_id":18029,"payload_sha256":"7ea50fc3af933a88f19115f17661e8e7c11be95ac0092b06bd8d2714540a0e16","signature_b64":"TtbdjuzC/5DXq2QquOdPQFDTJRKvMVSFaUZd5J7mByneTOnMBeFzQGRcOpkBE0jA4J/sFc54dJAjmSsAws9DBA==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-08-05T04:18:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RZvYAdnNIVk2n0a+KID8AvYmbUHpKDaYcVsYdqFOsxVQcuHIgQSEnWp3r8+w8fw24hJeHbwxNbz4uLrbJDHrBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T14:13:34.364000Z"},"content_sha256":"b3342bac0207afbd2ac2251918496643a6c9aa304a132ecdd746dc882c32f7f6","schema_version":"1.0","event_id":"sha256:b3342bac0207afbd2ac2251918496643a6c9aa304a132ecdd746dc882c32f7f6"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:NZ6UC77QVAQSZA2RNPTWACUPZ2","target":"integrity","payload":{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1186/s13040-023-00339-9) was visible in the surrounding text but could not be confirmed against doi.org as printed.","snippet":"Jesse G. Meyer, Ryan J. Urbanowicz, Patrick C. N. Martin, Karen O’Connor, Ruowang Li, Pei-Chen Peng, Tiffani J. Bright, Nicholas Tatonetti, Kyoung Jae Won, Graciela Gonzalez-Hernandez, and Jason H. Moore. 2023. ChatGPT and large language mo","arxiv_id":"2608.04006","detector":"doi_compliance","evidence":{"ref_index":44,"verdict_class":"incontrovertible","resolved_title":null,"printed_excerpt":"10.1186/s13040-","reconstructed_doi":"10.1186/s13040-023-00339-9"},"severity":"advisory","ref_index":44,"audited_at":"2026-08-05T04:18:05.690783Z","event_type":"pith.integrity.v1","detected_doi":"10.1186/s13040-023-00339-9","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"recoverable_identifier","evidence_hash":"33409c8bba956bafa975fd2b7b6089e28915cc51b22e196df1d5fea21c50900a","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.1.0","detected_arxiv_id":null,"integrity_event_id":18028,"payload_sha256":"236db8a7f3f17a29530a52652cd1cef7245bfac6ba00dfc20b6d335df514a114","signature_b64":"pK0pTfsDoPyaWWlAQIzc6aUovyb94DsJRpxFD/u3SL1n7QJopp53V0t31tpqePflV7xDCTRlZiS97BiodVbLDA==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-08-05T04:18:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"o/tiQcjNr68gpk6BVGwSvBxnKBZPF869znpG8a8q0hlfEH7lt50imWdYGlvy5lnMnwNLJBYy2xnJFrYau0SxAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T14:13:34.365336Z"},"content_sha256":"722905974c0f4fcb0d9cd52be0740bf4208b98f7eb0e22fc4dcf6d611b0869c8","schema_version":"1.0","event_id":"sha256:722905974c0f4fcb0d9cd52be0740bf4208b98f7eb0e22fc4dcf6d611b0869c8"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:NZ6UC77QVAQSZA2RNPTWACUPZ2","target":"integrity","payload":{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.18653/v1/2025.emnlp-main.1243) was visible in the surrounding text but could not be confirmed against doi.org as printed.","snippet":"Katerina Korre, Dimitris Tsirmpas, Nikos Gkoumas, Emma Cabalé, Danai Myrtzani, Theodoros Evgeniou, Ion Androut- sopoulos, and John Pavlopoulos. 2025. Evaluation and Facilitation of Online Discussions in the LLM Era: A Survey. In Proceedings","arxiv_id":"2608.04006","detector":"doi_compliance","evidence":{"ref_index":38,"verdict_class":"incontrovertible","resolved_title":null,"printed_excerpt":"10.18653/v1/2025.emnl","reconstructed_doi":"10.18653/v1/2025.emnlp-main.1243"},"severity":"advisory","ref_index":38,"audited_at":"2026-08-05T04:18:05.690783Z","event_type":"pith.integrity.v1","detected_doi":"10.18653/v1/2025.emnlp-main.1243","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"recoverable_identifier","evidence_hash":"a565bffc5e142261b44dd1185a9a7a35d7ec525f9009ba7ca91dd47dd35733e6","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.1.0","detected_arxiv_id":null,"integrity_event_id":18027,"payload_sha256":"2b8fcb74939e8a803acb504d19d90203474a585aa9b3862b385139e24893edc0","signature_b64":"5d5TFunD85M/7Vd2Efa3UttWTe9+qYCiU0jkFTB8Us52SNIQBicL82tSA3TftLWwQsTysTCAo6H9ZfUf1f8VDQ==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-08-05T04:18:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"SAjD++Ufx1T70ScujNlEVXAV01u1e75SRBCBxPDT1FeckWM35URAuNWsuP8NHjLOAtO/sI8kshLNSCnpIsmQBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T14:13:34.366076Z"},"content_sha256":"a6b417fc26eddd79cc09debcf5a282d22dca2cf57d0e060afd466454e5bd3c37","schema_version":"1.0","event_id":"sha256:a6b417fc26eddd79cc09debcf5a282d22dca2cf57d0e060afd466454e5bd3c37"},{"event_type":"integrity_finding","subject_pith_number":"pith:2026:NZ6UC77QVAQSZA2RNPTWACUPZ2","target":"integrity","payload":{"note":"DOI in the printed bibliography is fragmented by whitespace or line breaks. A longer candidate (10.1109/TVCG.2023.3346713) was visible in the surrounding text but could not be confirmed against doi.org as printed.","snippet":"Adam Coscia and Alex Endert. 2024. KnowledgeVIS: Interpreting Language Models by Comparing Fill-in-the-Blank Prompts.IEEE Transactions on Visualization and Computer Graphics30, 9 (2024), 6520–6532. doi:10.1109/TVCG.2023. 3346713","arxiv_id":"2608.04006","detector":"doi_compliance","evidence":{"ref_index":10,"verdict_class":"incontrovertible","resolved_title":null,"printed_excerpt":"10.1109/tvcg.2023","reconstructed_doi":"10.1109/TVCG.2023.3346713"},"severity":"advisory","ref_index":10,"audited_at":"2026-08-05T04:18:05.690783Z","event_type":"pith.integrity.v1","detected_doi":"10.1109/TVCG.2023.3346713","detector_url":"https://pith.science/pith-integrity-protocol#doi_compliance","external_url":null,"finding_type":"recoverable_identifier","evidence_hash":"ba665ed6406e669a1213fbf133fca5b061484b6e54d102aa07145351f7762dec","paper_version":1,"verdict_class":"incontrovertible","resolved_title":null,"detector_version":"1.1.0","detected_arxiv_id":null,"integrity_event_id":18026,"payload_sha256":"16d4fdd1954a48db226296d5d88d645477be8b05502b08383333fba9186966af","signature_b64":"CDP/IH1tB13SSaerHzy5CC04dxdyjz5dGh9iv2VNQnSc4N/7vy11JzMlFWDj4+eCLD+LVSaGwiTyY7zDgI8WCw==","signing_key_id":"pith-v1-2026-05"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-08-05T04:18:28Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QozyygPYoSS91QdlZn3fNzysM8/gIXPyObARleLYbibrfv1hVz9O/tYi6JUFbHMt36fDVUowxID52E11G70yAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-06T14:13:34.366733Z"},"content_sha256":"cf8f51a7c5ad88b9969931616530725bd9ce8c047672c43bc3d8bcfe1294e6d0","schema_version":"1.0","event_id":"sha256:cf8f51a7c5ad88b9969931616530725bd9ce8c047672c43bc3d8bcfe1294e6d0"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/NZ6UC77QVAQSZA2RNPTWACUPZ2/bundle.json","state_url":"https://pith.science/pith/NZ6UC77QVAQSZA2RNPTWACUPZ2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/NZ6UC77QVAQSZA2RNPTWACUPZ2/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":"2026-08-06T14:13:34Z","links":{"resolver":"https://pith.science/pith/NZ6UC77QVAQSZA2RNPTWACUPZ2","bundle":"https://pith.science/pith/NZ6UC77QVAQSZA2RNPTWACUPZ2/bundle.json","state":"https://pith.science/pith/NZ6UC77QVAQSZA2RNPTWACUPZ2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/NZ6UC77QVAQSZA2RNPTWACUPZ2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:NZ6UC77QVAQSZA2RNPTWACUPZ2","merge_version":"pith-open-graph-merge-v1","event_count":6,"valid_event_count":6,"invalid_event_count":0,"equivocation_count":1,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"18d665843be2eb4d206004ddbbeb53fa9ebb83c88e5873fb623b9205c8bf98c7","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-08-04T17:59:20Z","title_canon_sha256":"cc737e574cbf679df8661269a705d49992ab92c65a1c278eaf0dadb67a2b4b7e"},"schema_version":"1.0","source":{"id":"2608.04006","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2608.04006","created_at":"2026-08-05T01:37:41Z"},{"alias_kind":"arxiv_version","alias_value":"2608.04006v1","created_at":"2026-08-05T01:37:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2608.04006","created_at":"2026-08-05T01:37:41Z"},{"alias_kind":"pith_short_12","alias_value":"NZ6UC77QVAQS","created_at":"2026-08-05T01:37:41Z"},{"alias_kind":"pith_short_16","alias_value":"NZ6UC77QVAQSZA2R","created_at":"2026-08-05T01:37:41Z"},{"alias_kind":"pith_short_8","alias_value":"NZ6UC77Q","created_at":"2026-08-05T01:37:41Z"}],"graph_snapshots":[{"event_id":"sha256:8859e1090884d9b5215c57061d584980e139e20facb7e000a3c8ac97b187ab46","target":"graph","created_at":"2026-08-05T01:37:41Z","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/2608.04006/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"LLMs are reshaping educational technology, yet evaluating their responses for pedagogical alignment remains underexplored, relying heavily on the expertise of learning engineers building the technology. To bridge this gap, we explore trustworthiness as a structured lens for evaluation, leveraging existing measures of LLM trustworthiness to systematically identify potential pedagogical disruptions. Through a longitudinal co-design process with learning engineers developing an LLM-powered digital textbook, we: (1) co-constructed five trustworthiness metrics comprising 20 measures tailored to ped","authors_text":"Adam Coscia, Alex Endert, Langdon Holmes, Scott Crossley, Sujata Duwal","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-08-04T17:59:20Z","title":"Calibrating Trustworthiness: Co-Designing Metrics and Visualizations for Evaluating LLMs in Education"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2608.04006","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":"sha256:1b90c61caa227ca42a4bb51b2ae83798a6f7f5529dc2e14373c001eab4d1cfcb","target":"record","created_at":"2026-08-05T01:37:41Z","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":"18d665843be2eb4d206004ddbbeb53fa9ebb83c88e5873fb623b9205c8bf98c7","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.HC","submitted_at":"2026-08-04T17:59:20Z","title_canon_sha256":"cc737e574cbf679df8661269a705d49992ab92c65a1c278eaf0dadb67a2b4b7e"},"schema_version":"1.0","source":{"id":"2608.04006","kind":"arxiv","version":1}},"canonical_sha256":"6e7d417ff0a8212c83516be7600a8fce8f2b3d8846d29bfb967b9890cd48c9b6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6e7d417ff0a8212c83516be7600a8fce8f2b3d8846d29bfb967b9890cd48c9b6","first_computed_at":"2026-08-05T01:37:41.123137Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-08-05T01:37:41.123137Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"GGCt1263R0BTIy/xR5k/RTWNKm+ah1Dot9ebuYWsVPBzRciuQT0Jfv5sJWaLOjGt3/E0g/5bQFh3ggon/fKPCw==","signature_status":"signed_v1","signed_at":"2026-08-05T01:37:41.124630Z","signed_message":"canonical_sha256_bytes"},"source_id":"2608.04006","source_kind":"arxiv","source_version":1}}},"equivocations":[{"signer_id":"pith.science","event_type":"integrity_finding","target":"integrity","event_ids":["sha256:722905974c0f4fcb0d9cd52be0740bf4208b98f7eb0e22fc4dcf6d611b0869c8","sha256:a6b417fc26eddd79cc09debcf5a282d22dca2cf57d0e060afd466454e5bd3c37","sha256:b3342bac0207afbd2ac2251918496643a6c9aa304a132ecdd746dc882c32f7f6","sha256:cf8f51a7c5ad88b9969931616530725bd9ce8c047672c43bc3d8bcfe1294e6d0"]}],"invalid_events":[],"applied_event_ids":["sha256:1b90c61caa227ca42a4bb51b2ae83798a6f7f5529dc2e14373c001eab4d1cfcb","sha256:8859e1090884d9b5215c57061d584980e139e20facb7e000a3c8ac97b187ab46"],"state_sha256":"38b62b98f6a4ab482bbe7dde6fcf0268b454d9a97c38320e243a53f11c52dcd1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RGvVzHyVW0ohTMPFSZOIXvCye5edneCcIOZRqGqTZp8E1Hz6x7ErnCunmqMLfXfVQPDoUSGPtzJsqULjoenAAQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-06T14:13:34.383426Z","bundle_sha256":"4b8b62d28102cf8cca092d348656e110f9e5910048fd84f291d302708d5127a1"}}