Technology

base models

Pre-trained AI models on massive, diverse datasets: they serve as the versatile, general-purpose foundation for specialized applications.

Base models are the foundational core of modern AI: a single, large-scale model pre-trained on vast, diverse data. This paradigm shift (coined by Stanford HAI in 2021) allows for rapid deployment. Instead of building a bespoke model for every task, developers fine-tune a base model like OpenAI's GPT-4 or Google's BERT. This transfer learning approach drastically cuts development time and computational cost for downstream applications: a critical efficiency gain in the industry.

https://hai.stanford.edu/research/foundation-models

What builders pair with base models

Projects using both technologies. Select a pairing to see a project.

Built with base models

Agent 1 shared project Apertus-8B 1 shared project Claude Agent SDK 1 shared project FastAPI 1 shared project Google Notebook 1 shared project ImpliedAI 1 shared project

5 more pairings

LoRA 1 shared project PyTorch 1 shared project Qwen-2 1 shared project RAG 1 shared project Read 1 shared project

Pairing: Agent

Photo from the event — Event photo

The loss curve lied: catching hidden safety drift inside the fine-tuning, automated with an agent!

Zürich · July 1, 2026

view project | view Agent

Recent Talks & Demos

Showing 1-2 of 2

Members-Only

Sign in to see who built these projects

Sign in View FAQ

Agentic Safety Drift Detection

Claude Agent SDK Qwen-2