a reading list
Papers I love
Work I keep coming back to — for the idea, the framing, or the craft.
Favourites
paper · Cloud, Le et al., 2025
Subliminal Learning: Language Models Transmit Behavioral Traits via Hidden Signals in Data
Why this one stuck with you — a sentence or two. (placeholder — edit me)
paper · Fraser-Taliente, Kantamneni, Ong et al., 2026
Natural Language Autoencoders Produce Unsupervised Explanations of LLM Activations
What makes it worth reading. (placeholder — edit me)
paper · Jha, Zhang, Shmatikov, Morris, 2025
Harnessing the Universal Geometry of Embeddings
The idea you wish you'd thought of first. (placeholder — edit me)
paper · Huh, Cheung, Wang, Isola, 2024
The Platonic Representation Hypothesis
Why it shapes how you think about representations. (placeholder — edit me)