Recommendations
Explore
Contribute
About
Updates
Build
Fund
Research
Measure
AI & ML
Practice
Recommendations
Build
Fund
Research
Measure
AI & ML
Practice
Explore
Contribute
About
Updates
← Explore Research Questions
Research Question
What are the best methods for providing impartial robustness checking and critical friend support for output refinement?
Related Goals
Recommendations that are effectively stress tested to be as robust as possible.
Related Capabilities
Represent complexity
Substantiveness
Ability for final outputs to be nuanced, concrete, decisive, and comprehensive.
Related Existing Resources
Experimental Practice
Introducing Democratic Fine-Tuning
Joe Edelman’s Democratic Fine-Tuning (DFT) process aligning LLMs with human values through collective deliberation using Values Cards (where participants articulate underlying values like “protecting my community” rather than divisive language) and Moral Graphs (collaborative data structures mapp...
