← All incidents
✓ ResolvedConfidence: Low
Responses to fixed prompts no longer match this model’s own history (672 recent vs 2018 baseline responses). Biggest shifts: whether it reasoned at all on “Suggest a name for a pet cat. Reply with only the name.”; 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.”
Oct 8change began ≈ 6:05 PM UTC
Oct 10incident opened 12:38 AM UTC
Oct 10resolved 7:36 PM UTC
7.3e-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 10, 2026, 12:38 AM UTC UTC. Resolved Oct 10, 2026, 7:36 PM UTC UTC: the series returned to its baseline. Methodology.