Mastodawn

In a strict out-of-sample benchmark, our method improves real-time hallucination detection from ~64% to ~74%, with statistically decisive gains - even on a small, lightly trained model. Open to conversations with investors and partners in trustworthy AI
#TrustworthyAI #LLM

galen Feb 6

A new mechanism for the emergence of life from sparse beginnings:
"Crossing the Functional Desert: Cascade-Driven Assembly and Feasibility Transitions in Early Life"
https://arxiv.org/abs/2601.06272
#abiogenesis #functionalpercolation

galen Feb 6

Toward a physical theory of information processing:
"Functional Percolation: Criticality of Form and Function"
https://arxiv.org/abs/2512.09317
#complexsystems #complexityscience #networkscience #ai #percolation #cascades #criticality #artificialintelligence

galen Feb 25, 2025

https://sites.google.com/view/complexity-aiml/complexity-ai-ml

Complexity AI ML

Services

galen Oct 7, 2024

galen Sep 29, 2024

Just passed the initial round of the NSF SBIR application!

galen Sep 29, 2024

Just passed the initial round of the NSF SBIR application!

galen Aug 18, 2024

I’ve updated my GitHub projects gallery! http://galenwilkerson.github.io/
#ai #complexsystems #machinelearning

Galen J. Wilkerson: Insights in Machine Learning, AI, and Complex Systems Research.

galen Aug 7, 2024

CatHerder: Reflective LLM Code Creator and Debugger
https://lnkd.in/gWuP4FsP
CatHerder is a simple and intuitive code creator and debugger that utilizes Large Language Model (LLM) reflection to help you debug Python code effectively. By iteratively refining code with the assistance of an LLM, this tool aims to make the debugging process smoother and more efficient.

You can also use it to gradually add features to existing code.

#reflexion #LLM #LLMagents #agentic