SAVALERA
Research

Independent research on model behavior and human psychological safety

Our research focuses on how AI behaves in real-life, as well as how teams work with these systems in practice.

Open research docsDiscuss collaboration

Featured Documents

Savalera Lab Overview
Mission, goals, and roadmap for the research track.
Open document
Research Plan
Evaluation roadmap for language models and agents across behavior and performance.
Open document
Toxicity Echo Effect
Published paper on toxicity propagation in multi-turn LLM dialogues.
Open document

Explore or Collaborate

Follow publications and method releases, or discuss a focused collaboration where applied research can improve delivery quality.
Open research docsOpen newsTalk to us