Take audit: UfRn9bydpc-oSR8WFQ4co
2026-08-22T09:39:23.958Z — Mika
0/13 judged · keyboard: Space play/pause, ↑/↓ focus, P pass, T Translation failure, F fail
Do a blind check (transcripts hidden)
Ground Truth
What the Tester actually said. Never pre-filled from an Engine — type what you hear.
Transcripts
mini_prod_baseline gpt-4o-mini-transcribe · degraded · en · bare
— MER accuracy pending Ground Truth · 656 ms · verdict: unjudged
Teacher, I want to go to
mini_auto_bare_deg gpt-4o-mini-transcribe · degraded · auto · bare
— MER accuracy pending Ground Truth · 505 ms · verdict: unjudged
Teacher, I want to go to
mini_auto_bare_clean gpt-4o-mini-transcribe · clean · auto · bare
— MER accuracy pending Ground Truth · 693 ms · verdict: unjudged
這是耳光福高兔。
mini_auto_prompt_deg gpt-4o-mini-transcribe · degraded · auto · mixed prompt
— MER accuracy pending Ground Truth · 605 ms · verdict: unjudged
Teacher, I want to go to
mini_auto_prompt_clean gpt-4o-mini-transcribe · clean · auto · mixed prompt
— MER accuracy pending Ground Truth · 635 ms · verdict: unjudged
這是我往復購數。
mini_auto_prompt_64k gpt-4o-mini-transcribe · 64k · auto · mixed prompt
— MER accuracy pending Ground Truth · 606 ms · verdict: unjudged
Teacher, I want to go to
mini_zh_prompt_deg gpt-4o-mini-transcribe · degraded · zh · mixed prompt
— MER accuracy pending Ground Truth · 1172 ms · verdict: unjudged
Teacher, I want to go to
mini_zh_prompt_clean gpt-4o-mini-transcribe · clean · zh · mixed prompt
— MER accuracy pending Ground Truth · 684 ms · verdict: unjudged
這是我往復構圖。
gpt_tw_en_bare_deg gpt-transcribe · degraded · zh-tw + en · bare
— MER accuracy pending Ground Truth · 661 ms · verdict: unjudged
Teacher, I want to go too.
gpt_tw_en_prompt_deg gpt-transcribe · degraded · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 761 ms · verdict: unjudged
Teacher, I want to go too.
gpt_tw_en_prompt_clean gpt-transcribe · clean · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 622 ms · verdict: unjudged
Teacher, I want to go too.
gpt_tw_en_prompt_64k gpt-transcribe · 64k · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 761 ms · verdict: unjudged
Teacher, I want to go too.
gpt_tw_only_prompt_deg gpt-transcribe · degraded · zh-tw · mixed prompt
— MER accuracy pending Ground Truth · 1603 ms · verdict: unjudged
Teacher, I want to go too.
Winning configuration
Record the variant and rationale to hand to qa-rails after reviewing this Take.