The Normal Blog - System Two: The Normal Computing Blog

`posteriors`: Normal Computing’s library for Uncertainty-Aware LLMs

engineering

research

Scalable uncertainty quantification with PyTorch and Bayes

A First Demonstration of Thermodynamic Matrix Inversion

computing

research

We report on the first-ever experiment towards thermodynamic artificial intelligence: solving matrix inversion problems by allowing a system of coupled electrical oscillators to thermally equilibrate with its environment.

Denis Melanson, Max Hunter Gordon, Maxwell Aifer, Kaelan Donatella, Thomas Ahle, Gavin Crooks and Patrick J. Coles

Developing Advanced Reasoning and Planning Algorithms with LLMs

research

In this post we introduce Branches, our tool for prototyping and visualizing advanced LLM reasoning and planning algorithms. We apply Branches to the problem of generating Python code for HumanEval.

Arun Patro, Rami Sbahi, and Thomas Ahle

Infinite Context LLMs: Going Beyond RAG with Extended Minds

research

In this blog we discuss how the transformer architecture naturally extends over external memories, and share empirical results which leverage this capability to succeed where RAG has struggled. These methods are innate (don’t require fine tuning) and outperform popular retrieval augmented generation methods.

Phoebe Klett, Thomas Ahle

Explainable Language Models: Existing and Novel Approaches

research

We review key aspects of explainability for language models and introduce some Normal innovations.

Sam Duffield, Arun Patro, and Phoebe Klett

Eliminating hallucinations (fast!) in Large Language Models with Finite State Machines

engineering

In this blog, we introduce our method for regex-guided generation implemented in Outlines

Rémi Louf, Phoebe Klett, and Dan Simpson