Model components / Training & synthetic datasets

Stack-Edu

Hugging Face's education-filtered slice of StarCoder2Data, about 125 billion tokens of code scored by an educational-value classifier. It is a documented component of SmolLM3 and of the Dolma 3 pool behind OLMo 3.

The data terms defer to The Stack v2, which the card records. Verified 2026-08-13 via the HuggingFaceTB/stack-edu dataset card on Hugging Face.

Openness

4 medium confidence

4.0

access: ungated
license: inherits-the-stack-v2
dataset_card: present

Ungated download with a card; the card defers the data license to The Stack v2's terms.

https://huggingface.co/datasets/HuggingFaceTB/stack-edu recorded 2026-08-13
ungated access, card defers license to the-stack-v2

Adoption

2 high confidence

2.0

5,273 downloads in the trailing 30 days for HuggingFaceTB/stack-edu, which bands at 1K-10K, level 2 on the dataset adoption scale.

https://huggingface.co/api/datasets/HuggingFaceTB/stack-edu recorded 2026-08-13
5,273 downloads in the trailing 30 days for HuggingFaceTB/stack-edu

Capability

5 high confidence

5.0

Leads controlled classifier ablations, beating StarCoder2Data on MultiPL-E across all languages (SmolLM2 paper); current default in SmolLM3 and OLMo 3's Dolma 3.

https://arxiv.org/html/2502.02737v1 recorded 2026-08-13
Stack-Edu consistently outperforms StarCoder2Data on MultiPL-E (Python 25.6 vs 20.7, etc.)
https://huggingface.co/blog/smollm3 recorded 2026-08-13
Stack-Edu included in the SmolLM3 pretraining mixture

Verified 2026-08-13