Report
Opus 5.5 reportedly excels at labeling syllables from spectrograms but very often triggers safeguards
A user says the model extremely often triggers “reasoning extraction” safeguards while attempting a task they call it “insanely proficient” at.
TLDR
A user says Opus 5.5 is “insanely proficient” at labeling syllables from spectrograms, but extremely often triggers “reasoning extraction” safeguards while trying to do it.
Combined views
1.6K
2 Sources, first seen 4h ago
31 likes
