Full pipeline20260519T202149Z
Food Contaminants and Cross-Contamination
May 19, 2026, 20:21 UTC·17 ok · 7 skipped · 0 failed
- Language pairs
- 17
- Flagged segments
- 55
- of 74 scored
- Errors found
- 115
- 5461
- Corrected
- 32
- System xCOMET
- 0.5210.549
- triage mean · before ▸ after
- Threshold
- 0.80
- default
ModelsSegmentation: us.anthropic.claude-sonnet-4-6Scoring (xCOMET): xcomet-lite-endpoint (SageMaker)Error detection: us.anthropic.claude-opus-4-7Rewriting: us.anthropic.claude-opus-4-7
Manifest rebuilt from logs (original_pipeline_log, per_pair_pipeline_diff, es419_resume) — per-file counts are derived from the per-pair output.
en-es-419 — Originally failed on Stage 3 regex (es-419 digits); resumed and completed via --start-stage 3 after fix
Language pairs
| Severity | |||||
|---|---|---|---|---|---|
Amharicen-am | 4 | 7 | 34 | 2 | 0.5310.539▲+0.008 |
Burmeseen-my | 4 | 10 | 37 | 3 | 0.5570.572▲+0.016 |
Chinese (Simplified)en-zh-CN | 4 | 8 | 26 | 2 | 0.6850.706▲+0.021 |
Khmeren-km | 4 | 11 | 74 | 4 | 0.4160.515▲+0.099 |
Somalien-so | 4 | 11 | 65 | 1 | 0.2930.287▼-0.006 |
Swahili (Kenya)en-sw-KE | 4 | 7 | 52 | 2 | 0.4040.397▼-0.007 |
Swahili (Tanzania)en-sw-TZ | 4 | 10 | 64 | 2 | 0.4040.413▲+0.008 |
Vietnameseen-vi | 4 | 7 | 25 | 3 | 0.6680.698▲+0.030 |
Arabicen-ar | 3 | 6 | 24 | 2 | 0.5820.615▲+0.034 |
French (Canada)en-fr-CA | 3 | 3 | 21 | 2 | 0.7090.734▲+0.025 |
Hindien-hi | 3 | 5 | 23 | 3 | 0.5750.639▲+0.065 |
Portuguese (Brazil)en-pt-BR | 3 | 8 | 44 | 1 | 0.7330.742▲+0.009 |
Spanish (Latin America)en-es-419 | 3 | 6 | 42 | 1 | 0.6690.726▲+0.057 |
Tagalogen-tl | 3 | 7 | 34 | 3 | 0.4160.468▲+0.051 |
Ukrainianen-uk | 3 | 5 | 14 | 0 | 0.5340.534—0.000 |
Frenchen-fr | 2 | 4 | 22 | 1 | 0.6840.755▲+0.071 |
Englishen-en | 0 | 0 | — | 0 | 0.0000.000—0.000 |
Skipped — no xCOMET coverage
7HmonghmnHaitian CreolehtS'gaw KarenkswMarshallesemhOromoomKinyarwandarwTigrinyati
These target languages have no reliable xCOMET-lite metric, so the pipeline skipped them before scoring.