VoiceCodeBench: Evaluating Exact Structured-Token Recovery in Automatic Speech Recognition
Across systems, WER has little rank agreement with CTEM (Spearman $\rho=-0.28$) or TSR (Spearman $\rho=-0.28$), and even the strongest system leaves nearly one-third of recordings with an unrecovered critical value.
Tyler Baumgartner, Brandon Tai, Lisa Kaelin-Martin et al.
· 0 citations