← All incidents
✓ ResolvedConfidence: Low
Responses to fixed prompts no longer match this model’s own history (672 recent vs 3614 baseline responses). Biggest shifts: reasoning tokens on “Write one sentence about rain.”; visible output length on “Write a haiku about autumn.”; whether it reasoned at all on “Describe a cat in exactly five words.”
Oct 13change began ≈ 4:35 AM UTC
Oct 13incident opened 7:51 AM UTC
Oct 15resolved 10:13 AM UTC
2.4e-3adjusted 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 13, 2026, 7:51 AM UTC UTC. Resolved Oct 15, 2026, 10:13 AM UTC UTC: the series returned to its baseline. Methodology.