{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:7N6AP46ISHCN5W5TMZZVVOVNUM","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":"82b14c1cac3a1870799bf3edc1dc12f96909aef0028924cae8e847494da8ba0e","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T07:06:34Z","title_canon_sha256":"23fc7fdf433f5ea7df2b4fd0933b5c1fdc7c048127bff01ed33683024c09565a"},"schema_version":"1.0","source":{"id":"2605.24904","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.24904","created_at":"2026-05-26T01:04:04Z"},{"alias_kind":"arxiv_version","alias_value":"2605.24904v1","created_at":"2026-05-26T01:04:04Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.24904","created_at":"2026-05-26T01:04:04Z"},{"alias_kind":"pith_short_12","alias_value":"7N6AP46ISHCN","created_at":"2026-05-26T01:04:04Z"},{"alias_kind":"pith_short_16","alias_value":"7N6AP46ISHCN5W5T","created_at":"2026-05-26T01:04:04Z"},{"alias_kind":"pith_short_8","alias_value":"7N6AP46I","created_at":"2026-05-26T01:04:04Z"}],"graph_snapshots":[{"event_id":"sha256:8844ae557dd3847cc15a4c950b182f4cc938bbc7e7ac8134e405b1c1f74caadb","target":"graph","created_at":"2026-05-26T01:04:04Z","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/2605.24904/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Machine-translated benchmarks are widely used to assess the multilingual capabilities of large language models (LLMs), yet translation errors in these benchmarks remain underexplored, raising concerns about the reliability and comparability of multilingual evaluation. We address two practical gaps: (i) how well automatic MQM-style error spans from LLM judges and a span-aware QE baseline (xCOMET-XXL) match expert human span annotations on benchmark translations, and (ii) how strongly translation errors (as opposed to source-side issues in the English original) explain accuracy drops on translat","authors_text":"Bernhard Stadler, Jens Lehmann, Klaudia-Doris Thellmann, Michael F\\\"arber","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T07:06:34Z","title":"Quantifying the Impact of Translation Errors on Multilingual LLM Evaluation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.24904","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:b6ec5f908ce5b6d38adb54464341cba9cc681966a6bb4c7c3f0f96c55b678dd1","target":"record","created_at":"2026-05-26T01:04:04Z","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":"82b14c1cac3a1870799bf3edc1dc12f96909aef0028924cae8e847494da8ba0e","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2026-05-24T07:06:34Z","title_canon_sha256":"23fc7fdf433f5ea7df2b4fd0933b5c1fdc7c048127bff01ed33683024c09565a"},"schema_version":"1.0","source":{"id":"2605.24904","kind":"arxiv","version":1}},"canonical_sha256":"fb7c07f3c891c4dedbb366735abaada32b53856123ca1890d73228fcfe01e33e","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"fb7c07f3c891c4dedbb366735abaada32b53856123ca1890d73228fcfe01e33e","first_computed_at":"2026-05-26T01:04:04.702414Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-26T01:04:04.702414Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"LVROUm2TX+9VueCneAddz4bzWsrJyI8Dmih+K4QKyaWGRK36Txc+7xILGGevTbet5dQS3PkHS9MTAZYuiLhmCQ==","signature_status":"signed_v1","signed_at":"2026-05-26T01:04:04.702992Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.24904","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b6ec5f908ce5b6d38adb54464341cba9cc681966a6bb4c7c3f0f96c55b678dd1","sha256:8844ae557dd3847cc15a4c950b182f4cc938bbc7e7ac8134e405b1c1f74caadb"],"state_sha256":"953e7f48739ff72bc75d0e94761b9bf8febb14d03bd6d9d21a19eed43dcd1388"}