dvyomkesh/

opsd_edge_dataset

Edge prompts for On-Policy Self-Distillation training on Nemotron-3-Nano.

"Self-Distilled Reasoner: On-Policy Self-Distillation for LLMs"

Paper: arXiv:2601.18734 Code: github.com/siyan-zhao/OPSD

This dataset contains 1,782 "edge" prompts where the 0.83 Nemotron adapter achieves 25-75% pass rate (uncertain cases ideal for learning).

From Figure 2 of the OPSD… See the full description on the dataset page: https://huggingface.co/datasets/dvyomkesh/opsd_edge_dataset.