This is your work, valued

Leon Chlon

Elite
@leochlon

My research interests focus on machine learning systems, AI safety, and mechanistic understanding of modern foundation models, particularly around hallucination

hallbayes. Python

1.7k

ntkmirror. Python

347

mezzanine. Python

145

ITO. Python

65

teta. TeX

36

factuality-slice. Python

30

openbook. TeX

22

berry. Python

15

medium. Jupyter Notebook

9

bayesian-flashcards. A modern, extensible, and research-backed flashcard app designed for optimal learning, focus, and memory—especially for users with ADHD and anyone who wants smarter, science-based study sessions.

6

aecf. Code accompanying Adaptive Entropy-Gated Contrastive Fusion: Robust Multimodal Learning under Missing Modalities

5

procedural_chlon2026. Supporting experiments

3

triton-blackwell. Python

2

bgz.

1

optimistix. Nonlinear optimisation (root-finding, least squares, ...) in JAX+Equinox. https://docs.kidger.site/optimistix/

1

cvhack. Python

1

pytorch. Tensors and Dynamic neural networks in Python with strong GPU acceleration

1

optax. Optax is a gradient processing and optimization library for JAX.

1

rfcs. PyTorch RFCs (experimental)

1

transformers. 🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

1

leochlon.github.io. JavaScript

1

watermark-edfl-toolkit. Information-theoretic robustness certification toolkit for logits-based text watermarks (SynthID, KGW-family).

1