← All incidents
● OpenConfidence: Low
Responses to fixed prompts no longer match this model’s own history (1344 recent vs 7961 baseline responses). Biggest shifts: output length on “Pick a day of the week. Reply with one word.”; reasoning tokens on “Pick a day of the week. Reply with one word.”; whether it reasoned at all on “all prompts together”
Oct 12change began ≈ 4:40 PM UTC
Oct 14incident opened 6:41 AM UTC
—still open
6.7e-4adjusted p
What this means
Raven Alpha compares how the model answers a fixed set of open-ended prompts with its own history. A persistent, significant shift means the model, its configuration or its serving path changed, even when every scored test still passes.
Opened Oct 14, 2026, 6:41 AM UTC UTC. An incident resolves on its own when the series returns to its baseline. Methodology.