datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
italic-softkd-pool
italic-softkd-pool
The exact training data of idealab-cs2/zagreus-0.4B-italic-softkd: 21,606 Italian multiple-choice questions with committee soft labels. One soft-KD training run from mii-llm/zagreus-0.4B-ita on the train split reaches 0.4787 on the full ITALIC 10K (official harness, 5-shot fast, temperature 0), from a 0.2802 base.
train is the full pool; the other three splits partition it by provenance:
split
rows
contents
train
21,606
the full training file (union… See the full description on the dataset page: https://huggingface.co/datasets/idealab-cs2/italic-softkd-pool.italic-extkd-pool
italic-extkd-pool
The stage-3 training data of idealab-cs2/zagreus-0.4B-italic-extkd: 57,563 Italian multiple-choice questions from public, in-distribution datasets with teacher soft labels. One soft-KD stage from the stage-2 checkpoint on the agreement-filtered subset (28,561 items where the teacher agrees with the gold answer) reaches 0.4921 / 0.4929 / 0.4932 on the full ITALIC 10K (official harness, 5-shot fast, temperature 0, three independent runs). Full lineage: 0.2802… See the full description on the dataset page: https://huggingface.co/datasets/idealab-cs2/italic-extkd-pool.