@misc{indiciae83c247bbe77b,
  title = {Predictive Concept Decoders: Training Scalable End-to-End Interpretability Assistants},
  author = {Vincent Huang and Dami Choi and Daniel D. Johnson and Sarah Schwettmann and Jacob Steinhardt},
  year = {2025},
  url = {https://arxiv.org/abs/2512.15712},
  note = {Source identifier: 2512.15712}
}