New AI Alignment Hub Centralizes 120+ Reported Incidents of Model MisbehaviorAIElie BakouchThomas WolfOct 26