{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:OEKNMEPMS55KJWQJK4WRI7SWPO","short_pith_number":"pith:OEKNMEPM","canonical_record":{"source":{"id":"2601.22025","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-29T17:32:34Z","cross_cats_sorted":["cs.AI","cs.IR","cs.SE"],"title_canon_sha256":"6165e6e7f0192571d3d226663d9cd45d7d8364344e8ccc2724da570aed748c70","abstract_canon_sha256":"bf58489e53a2cb81f196da61dbc3e63d71786a0a1db90e1723fde69e412661c7"},"schema_version":"1.0"},"canonical_sha256":"7114d611ec977aa4da09572d147e567b9fc4592ddd877c6830a57ead0a56ee29","source":{"kind":"arxiv","id":"2601.22025","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2601.22025","created_at":"2026-06-11T01:09:27Z"},{"alias_kind":"arxiv_version","alias_value":"2601.22025v2","created_at":"2026-06-11T01:09:27Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.22025","created_at":"2026-06-11T01:09:27Z"},{"alias_kind":"pith_short_12","alias_value":"OEKNMEPMS55K","created_at":"2026-06-11T01:09:27Z"},{"alias_kind":"pith_short_16","alias_value":"OEKNMEPMS55KJWQJ","created_at":"2026-06-11T01:09:27Z"},{"alias_kind":"pith_short_8","alias_value":"OEKNMEPM","created_at":"2026-06-11T01:09:27Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:OEKNMEPMS55KJWQJK4WRI7SWPO","target":"record","payload":{"canonical_record":{"source":{"id":"2601.22025","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-29T17:32:34Z","cross_cats_sorted":["cs.AI","cs.IR","cs.SE"],"title_canon_sha256":"6165e6e7f0192571d3d226663d9cd45d7d8364344e8ccc2724da570aed748c70","abstract_canon_sha256":"bf58489e53a2cb81f196da61dbc3e63d71786a0a1db90e1723fde69e412661c7"},"schema_version":"1.0"},"canonical_sha256":"7114d611ec977aa4da09572d147e567b9fc4592ddd877c6830a57ead0a56ee29","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-11T01:09:27.793066Z","signature_b64":"TGWhUQRPvzXru34DHp0qIacXv3fnIuYDxNwfMF7leQj1ZUC57sRMrw4CnWJihwDroVE+kKJFeF+TDPci7oVnDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7114d611ec977aa4da09572d147e567b9fc4592ddd877c6830a57ead0a56ee29","last_reissued_at":"2026-06-11T01:09:27.792054Z","signature_status":"signed_v1","first_computed_at":"2026-06-11T01:09:27.792054Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2601.22025","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-06-11T01:09:27Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"OS81W2/PQMvwek+yOcSUkk5g+1h7qkkjaVGLO6GzCDcfxz2M175//qu3AVomZzlUWEV4sh0hGO13MiYyw6RVDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T10:27:20.606514Z"},"content_sha256":"a50a46448a1f38549d02ceb02d4d546a8569bb94278dc2579e316b2006a15dad","schema_version":"1.0","event_id":"sha256:a50a46448a1f38549d02ceb02d4d546a8569bb94278dc2579e316b2006a15dad"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:OEKNMEPMS55KJWQJK4WRI7SWPO","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"When Generic Prompt Improvements Hurt: Evaluation-Driven Iteration for LLM Applications","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.IR","cs.SE"],"primary_cat":"cs.CL","authors_text":"Daniel Commey","submitted_at":"2026-01-29T17:32:34Z","abstract_excerpt":"Evaluating Large Language Model (LLM) applications differs from conventional software testing because outputs are probabilistic, semantically variable, and sensitive to prompt and model changes. This technical report proposes the Minimum Viable Evaluation Suite (MVES), an audit-oriented structure for application-level LLM evaluation. MVES links application categories to failure modes, metrics, required artifacts, and validation evidence across general LLM applications, retrieval-augmented systems, and agentic workflows. We pair the framework with a reproducible local evaluation harness coverin"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.22025","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/2601.22025/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":"2026-06-11T01:09:27Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"nysW4bjVuHCwmgfAgzViAEoUXXt24JKAf7aohW7K1RRckygmKCXgDXKYEwSep3EPE1FN5nz9aKQdqZ0d0bqLDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-03T10:27:20.606913Z"},"content_sha256":"96475fa033760dd890e24d9972f2b826ca2f244fee5b25daeefa064b00604a15","schema_version":"1.0","event_id":"sha256:96475fa033760dd890e24d9972f2b826ca2f244fee5b25daeefa064b00604a15"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/OEKNMEPMS55KJWQJK4WRI7SWPO/bundle.json","state_url":"https://pith.science/pith/OEKNMEPMS55KJWQJK4WRI7SWPO/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/OEKNMEPMS55KJWQJK4WRI7SWPO/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":"2026-08-03T10:27:20Z","links":{"resolver":"https://pith.science/pith/OEKNMEPMS55KJWQJK4WRI7SWPO","bundle":"https://pith.science/pith/OEKNMEPMS55KJWQJK4WRI7SWPO/bundle.json","state":"https://pith.science/pith/OEKNMEPMS55KJWQJK4WRI7SWPO/state.json","well_known_bundle":"https://pith.science/.well-known/pith/OEKNMEPMS55KJWQJK4WRI7SWPO/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:OEKNMEPMS55KJWQJK4WRI7SWPO","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":"bf58489e53a2cb81f196da61dbc3e63d71786a0a1db90e1723fde69e412661c7","cross_cats_sorted":["cs.AI","cs.IR","cs.SE"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-29T17:32:34Z","title_canon_sha256":"6165e6e7f0192571d3d226663d9cd45d7d8364344e8ccc2724da570aed748c70"},"schema_version":"1.0","source":{"id":"2601.22025","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2601.22025","created_at":"2026-06-11T01:09:27Z"},{"alias_kind":"arxiv_version","alias_value":"2601.22025v2","created_at":"2026-06-11T01:09:27Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2601.22025","created_at":"2026-06-11T01:09:27Z"},{"alias_kind":"pith_short_12","alias_value":"OEKNMEPMS55K","created_at":"2026-06-11T01:09:27Z"},{"alias_kind":"pith_short_16","alias_value":"OEKNMEPMS55KJWQJ","created_at":"2026-06-11T01:09:27Z"},{"alias_kind":"pith_short_8","alias_value":"OEKNMEPM","created_at":"2026-06-11T01:09:27Z"}],"graph_snapshots":[{"event_id":"sha256:96475fa033760dd890e24d9972f2b826ca2f244fee5b25daeefa064b00604a15","target":"graph","created_at":"2026-06-11T01:09:27Z","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/2601.22025/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Evaluating Large Language Model (LLM) applications differs from conventional software testing because outputs are probabilistic, semantically variable, and sensitive to prompt and model changes. This technical report proposes the Minimum Viable Evaluation Suite (MVES), an audit-oriented structure for application-level LLM evaluation. MVES links application categories to failure modes, metrics, required artifacts, and validation evidence across general LLM applications, retrieval-augmented systems, and agentic workflows. We pair the framework with a reproducible local evaluation harness coverin","authors_text":"Daniel Commey","cross_cats":["cs.AI","cs.IR","cs.SE"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-29T17:32:34Z","title":"When Generic Prompt Improvements Hurt: Evaluation-Driven Iteration for LLM Applications"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2601.22025","kind":"arxiv","version":2},"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:a50a46448a1f38549d02ceb02d4d546a8569bb94278dc2579e316b2006a15dad","target":"record","created_at":"2026-06-11T01:09:27Z","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":"bf58489e53a2cb81f196da61dbc3e63d71786a0a1db90e1723fde69e412661c7","cross_cats_sorted":["cs.AI","cs.IR","cs.SE"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-01-29T17:32:34Z","title_canon_sha256":"6165e6e7f0192571d3d226663d9cd45d7d8364344e8ccc2724da570aed748c70"},"schema_version":"1.0","source":{"id":"2601.22025","kind":"arxiv","version":2}},"canonical_sha256":"7114d611ec977aa4da09572d147e567b9fc4592ddd877c6830a57ead0a56ee29","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"7114d611ec977aa4da09572d147e567b9fc4592ddd877c6830a57ead0a56ee29","first_computed_at":"2026-06-11T01:09:27.792054Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-11T01:09:27.792054Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"TGWhUQRPvzXru34DHp0qIacXv3fnIuYDxNwfMF7leQj1ZUC57sRMrw4CnWJihwDroVE+kKJFeF+TDPci7oVnDg==","signature_status":"signed_v1","signed_at":"2026-06-11T01:09:27.793066Z","signed_message":"canonical_sha256_bytes"},"source_id":"2601.22025","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:a50a46448a1f38549d02ceb02d4d546a8569bb94278dc2579e316b2006a15dad","sha256:96475fa033760dd890e24d9972f2b826ca2f244fee5b25daeefa064b00604a15"],"state_sha256":"0526a43584c8d0f74eedf05d0b0c1b8f5c6818150b75310584402e80fd1d53b3"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"5pCXUSLVJxmiHElk6g6u58mAPfvLXkXM2RHtOO3aRghJRBYAF0PcNZw/FrZAP4crJyBSX6JP6jogMzSMUOahDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-03T10:27:20.610261Z","bundle_sha256":"b5863ae1293aecaece173a91cb97b88df93b6305ce988fa2d5aff78d8f36edc1"}}