Meet your research partner for AI’s hardest questions at NeurIPS
Trusted by researchers at the frontier
Make room for a better conversation.
Our research at NeurIPS

MetaLoop: Benchmarking the Full Metacognitive Loop in LLMs
Thursday, Dec 10 at 17:00 in Hall 1-4 (Sydney Poster Session 6)
A seven-task benchmark that tests how models monitor uncertainty, act on it and learn from feedback. The research compares 12 frontier models with 460 human participants—and finds that accuracy does not predict metacognitive ability.
Nora Petrova, John Burden and Jerome Wynne

Theory of Mind and Persuasion Beyond Conversation
Tuesday, Dec 8 at 10:00 in Hall 1-4 (Sydney Poster Session 1)
Can a model shape another agent’s beliefs through planning and action? This work tests social reasoning beyond dialogue, with implications for how we evaluate autonomous agents.
Led by Ben Slater, with Matteo G. Mecattaf, Lucy G. Cheke, John Burden and Winnie Street.
NeurIPS 2026 Workshop: Dynamic Alignment in Human–AI Coupled Systems
Our team has an accepted paper at the BiAlign workshop. Come along to explore the research, bring your questions and join the discussion in Sydney.
More from the NeurIPS community.
Accepted at NeurIPS 2026, these papers cite Prolific and were authored by independent research teams outside Prolific.
Working through a complex research question?
Bring a difficult research question, an evaluation that is not telling you enough, or a study you are unsure how to run. A draft protocol or example task is useful, but you do not need a finished brief. Tell us what you want to learn and where you are getting stuck.
Start with the question you need to answer. Our research consultants can support study design and delivery, with managed support for multi-stage studies and bespoke workflows. Together, we can scope the part of your programme where additional methodological or operational support would make the greatest difference.
Yes. Design guidance is part of our managed research support. Bring the hypothesis, the decision your evidence needs to inform and the uncertainties in your approach. That gives us a starting point for discussing the study design, practical constraints and an appropriate first phase.
Relevant expertise is a starting point. Prolific supports specialist qualification, credential verification and task calibration. For your study, we can discuss what evaluators need to know, what good judgment looks like and how to assess readiness before the main evaluation.
Disagreement can reveal ambiguous instructions, differences in expertise or legitimate differences in perspective. It deserves investigation. Bring an example to our conversation: we can discuss what agreement should mean for your research question and where a pilot could help test the approach.
Our research services include longitudinal and multi-phase study consultancy, participant tracking and retention support. We can discuss cohort continuity, changes to the protocol and the practical demands of each phase. Recontact depends on participant consent and availability.
Depending on the agreed scope, managed support can cover setup, evaluator sourcing, onboarding, calibration, quality checks and delivery coordination. Your team retains scientific direction. We would agree responsibilities, review points and deliverables together before work begins.
Yes. Human evaluation can connect to your workflow through the API, while bespoke research support can include custom tooling and integrations. Share your current setup and the steps that slow the research down so we can discuss where support would be useful.




