Tag: interpretability

Showing 1-2 of 2

Feb 06, 20267 min readAgent PlaybookIntermediate5 min build

InterPReT: Interactive policy restructuring enables laypersons to train more robust imitation policies

InterPReT lets lay users restructure a policy via instructions and continue training from demonstrations. In a 34-user racing game study it improved robustness without hurting usability.

imitation-learning human-in-the-loop interactive-training hci interpretability

+2 more

ml-systems arxiv-2602.04213

NewsUnited Kingdom

Open

Dec 16, 20257 min readModel Release BriefIntermediate120 min build

Gemma Scope 2 expands open interpretability and reproducible traces across the Gemma 3 family

Gemma Scope 2 makes open interpretability tools and reproducible trace exports available across the Gemma 3 family, enabling safety teams to probe and audit complex LLM behavior.

Gemma Scope 2 Gemma 3 DeepMind interpretability AI safety

+3 more

open tools ML engineering founder