research paper

A position paper arguing that AI alignment methods can be repurposed for censorship and manipulation, with examples across pre-training, post-training, and inference-time controls.

Apple Machine Learning Research paper proposing LaDiR, a reasoning framework that combines a VAE-based latent space with latent diffusion to improve LLM text reasoning and iterative refinement.

An ICML 2026 research project page arguing that prompt injection comes from how LLMs misread roles, with an extended writeup, examples, and links to the paper, code, arXiv, and BibTeX.

Research paper introducing hyperagents, a self-referential agent framework that combines a task agent and a meta agent into one editable program. The abstract describes a DGM-based system that improves both task performance and its own improvement process across domains.

A DeepSeek paper about DSpark, a full-stack codebase for training and evaluating speculative decoding algorithms to speed up LLM inference.