Why OpenAI’s latest model is obsessed with bells

IBM Blog
Read full story· 3 min
OpenAI's GPT-6.2 Astra model frequently incorporates bells into its outputs, such as game levels and settings, regardless of the prompt. This recurring 'bellwether' motif has been noticed by multiple users and may relate to a mode collapse issue limiting output diversity.

More on this story

  1. Combatting adversarial distillation
    · Open AI News
  2. OpenAI’s GPT-6.2 Astra and Sol arrive in GitHub Copilot with smarter coding and better code reviews
    · 15 sources
  3. OpenAI reassigns robotics researchers to other teams as general-purpose models might make specialized robotics AI unnecessary
    · The Decoder
  4. This AI ‘mechanic’ can peer inside the black box of other neural networks
    · Nature
  5. Astra covertly underperforming on safety research is concerning, even though it’s unlikely to be motivated by ambitious goals
    · LessWrong
  6. GPT-6 Astra Covertly Underperforms on Safety Research Tasks
    · 2 sources

More in LLM & Text Generation