01Evaluation
Benchmarks and Evaluation
We develop the science of LLM evaluation, setting the standard for rigorous assessment and identifying hidden risks before they matter.
ExploreOxford Internet Institute · LabEst. 2022 · Oxford, United Kingdom
Advancing the science of AI evaluation, safety, and reasoning — to improve how machines and humans decide.
§ 01Research
Three lines, asking whether we can trust what these systems do next.
02Safety
AI Safety and Security
From bias and toxicity to agentic misalignment, we study the full spectrum of AI risk and develop the technical and governance tools to address it.
Explore03Agentic
Agentic AI for Science
We build agentic systems that automate scientific knowledge synthesis and discovery, with a focus on agents that are reliable, transparent and domain-grounded.
Explore§ 03Engage
Four ways to work with the lab.
We're selective by design. Each track has its own intake cadence — pick the one closest to how you work.
01Researchers & postdocs
Collaborate on evaluation methodology, safety benchmarks, and agentic-science platforms. We host visiting researchers and run joint projects with peer labs across Europe, North America, and Asia.
Visit or collaborate02Industry partners
Workshops, co-built evaluation harnesses, and multi-year research partnerships for teams shipping LLM products where reliability matters — healthcare, finance, retail, and government.
Engage with OxRML03Policymakers
Evidence-based briefings on evaluation, model risk, and agentic deployment. We translate empirical findings into the questions regulators are actually asking.
See our policy work04MSc Social Data Science
A one-year master’s at the Oxford Internet Institute, combining the social sciences with computational methods. Applications are typically due in early January for September entry.
MSc programme05DPhil Social Data Science
Apply for fully-funded DPhil studentships at the Oxford Internet Institute. We supervise across evaluation, AI safety, agentic systems, and human–AI interaction. Applications are typically due in early January for September entry.
DPhil programme§ 04Latest at OxRML
What we've been shipping.
ConferenceApril 2026
OxRML at ICLR 2026
AwardFebruary 2026
Ryan Othniel Kearns Wins MSc Thesis Prize
ResearchFebruary 2026