Safety Concerns
Connor and Daniel discuss the safety concerns surrounding the release of GPT-2 and the decision not to release the model based on the potential precedent it could set. They explore the idea of solidarity and the importance of considering the impact of releasing potentially dangerous models.In this clip
From this podcast
The Gradient
Connor Leahy on EleutherAI, Replicating GPT-2/GPT-3, AI Risk and Alignment
Related Questions