@misc{indiciae4edbbf9e4163,
  title = {The Neutral Mask: How Alignment Training Provides Shallow Alignment while Leaving Partisan Structure Intact in a Large Language Model},
  author = {Wendy K. Tam},
  year = {2026},
  url = {https://arxiv.org/abs/2606.09735},
  note = {Source identifier: 2606.09735}
}