←

unified capability-based benchmark for AI memory systems

first time writing benchmarks so not the best work but since then learned a lot about how to design task-based vs. metric-based benchmarks, how to select datasets, and how to ensure reproducibility.

view the GitHub repository ↗