Guide Labs Debuts Steerling-8B: The First Truly Interpretable LLM

A San Francisco startup is solving one of AI biggest problems: understanding why large language models do what they do.

Guide Labs, founded by CEO Julius Adebayo and chief science officer Aya Abdelsalam Ismail, has open-sourced Steerling-8B—an 8-billion-parameter LLM with a novel architecture that makes every token traceable back to its origins in the training data.

The Interpretability Problem

Current large language models are essentially black boxes. Whether it is xAI struggling to fine-tune Grok political biases, ChatGPT issues with sycophancy, or run-of-the-mill hallucinations, the fundamental challenge remains: plumbing through a neural network with billions of parameters is not easy.

How Steerling Works

Guide Labs flipped the script on interpretability research. Instead of doing neuroscience on a model after training, they engineered interpretability from the ground up.

The key innovation: A concept layer that buckets data into traceable categories during training.

Performance Claims

Guide Labs says Steerling-8B achieves 90% of the capability of existing frontier models while using less training data—thanks to its novel architecture.

“This model demonstrates that training interpretable models is no longer a sort of science; it is now an engineering problem,” Adebayo said.

Company Background

• Founded: By Julius Adebayo (CEO) and Aya Abdelsalam Ismail (CSO)
• Origin: Emerged from Y Combinator
• Funding: $9 million seed round from Initialized Capital (November 2024)

Key Takeaways

• Steerling-8B: 8B parameter LLM with built-in interpretability
• Novel architecture: Concept layer makes every token traceable
• Performance: 90% of frontier models with less training data
• Use cases: Regulated industries, scientific research, consumer safety

As AI systems become more powerful, interpretability transitions from a nice-to-have to a necessity.