Take audit: PSDfF_KTVjRBYPvDtTG7D
2026-08-22T09:13:01.884Z — 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 · 577 ms · verdict: unjudged
我望這個to
mini_auto_bare_deg gpt-4o-mini-transcribe · degraded · auto · bare
— MER accuracy pending Ground Truth · 650 ms · verdict: unjudged
我望這個to
mini_auto_bare_clean gpt-4o-mini-transcribe · clean · auto · bare
— MER accuracy pending Ground Truth · 692 ms · verdict: unjudged
我望著這個to
mini_auto_prompt_deg gpt-4o-mini-transcribe · degraded · auto · mixed prompt
— MER accuracy pending Ground Truth · 670 ms · verdict: unjudged
我want這個to
mini_auto_prompt_clean gpt-4o-mini-transcribe · clean · auto · mixed prompt
— MER accuracy pending Ground Truth · 986 ms · verdict: unjudged
我望著這個to
mini_auto_prompt_64k gpt-4o-mini-transcribe · 64k · auto · mixed prompt
— MER accuracy pending Ground Truth · 1024 ms · verdict: unjudged
我望著這個to
mini_zh_prompt_deg gpt-4o-mini-transcribe · degraded · zh · mixed prompt
— MER accuracy pending Ground Truth · 1011 ms · verdict: unjudged
我望著這個to
mini_zh_prompt_clean gpt-4o-mini-transcribe · clean · zh · mixed prompt
— MER accuracy pending Ground Truth · 957 ms · verdict: unjudged
我望成這個 to
gpt_tw_en_bare_deg gpt-transcribe · degraded · zh-tw + en · bare
— MER accuracy pending Ground Truth · 576 ms · verdict: unjudged
我忘這個 to。
gpt_tw_en_prompt_deg gpt-transcribe · degraded · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 594 ms · verdict: unjudged
我忘了這個 to。
gpt_tw_en_prompt_clean gpt-transcribe · clean · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 544 ms · verdict: unjudged
我 want 這個 to。
gpt_tw_en_prompt_64k gpt-transcribe · 64k · zh-tw + en · mixed prompt
— MER accuracy pending Ground Truth · 739 ms · verdict: unjudged
我忘了這個 to。
gpt_tw_only_prompt_deg gpt-transcribe · degraded · zh-tw · mixed prompt
— MER accuracy pending Ground Truth · 593 ms · verdict: unjudged
我忘了這個 to。
Winning configuration
Record the variant and rationale to hand to qa-rails after reviewing this Take.