arjhinety/small-mind-pmb-v0
PMB v0 — Personalised Memory Benchmark An evaluation benchmark for long-horizon personalised memory in small language models. It asks whether a model can recall what a specific user told it across many sessions, and — the part most memory benchmarks skip — whether it can decline to answer when the memory does not contain the answer. Built for small-mind-companion, a study of how much of the long-horizon memory gap a ~2B multimodal model can close without scaling parameters. Part… See the full description on the dataset page: https://huggingface.co/datasets/arjhinety/small-mind-pmb-v0.
Correct the card to the nested layout and both pin recipes
Publish PMB v0 with the source layout and both integrity pins
Restructure to mirror the source layout so both hash.txt pins are recomputable
PMB v0 and the Study 001 probe sets
initial commit
