tech
February 23, 2026
Guide Labs debuts a new kind of interpretable LLM
The company open-sourced an 8 billion parameter LLM, Steerling-8B, trained with a new architecture designed to make its actions easily interpretable.

TL;DR
- Guide Labs open-sourced Steerling-8B, an 8 billion parameter LLM with a new, interpretable architecture.
- Every token produced by Steerling-8B can be traced back to its origins in the training data.
- This architecture aims to make AI actions easily understandable, addressing issues like hallucinations and bias.
- The interpretability is engineered into the model from the ground up, rather than requiring post-hoc analysis.
- Steerling-8B retains emergent behaviors and discovered concepts, according to CEO Julius Adebayo.
- The company believes this interpretable architecture is essential for consumer-facing LLMs, regulated industries, and scientific applications.
- Guide Labs claims Steerling-8B achieves 90% of the capability of existing models with less training data.
- The company plans to build a larger model and offer API and agentic access.
Continue reading the original article