Model components / Training & synthetic datasets

UltraChat 200k

UltraChat 200k is Hugging Face H4's filtered 200k-dialogue slice of the ChatGPT-distilled UltraChat corpus, released under MIT for supervised fine-tuning. It anchored the Zephyr recipe and became a default open SFT set for chat alignment, also used by TinyLlama-Chat.

Openness

5 high confidence

5.0

license: MIT
gated: false
dataset_card: present

MIT-licensed, ungated, full dataset card.

https://huggingface.co/datasets/HuggingFaceH4/ultrachat_200k recorded 2026-07-04
MIT license, ungated access, dataset card

Adoption

3 high confidence

3.0

57,823 monthly downloads on Hugging Face, graded on the training-corpus bands.

https://huggingface.co/api/datasets/HuggingFaceH4/ultrachat_200k recorded 2026-07-04
downloads field = 57,823 (30-day)

Capability

3 high confidence

3.0

Clean SFT before/after in the Zephyr paper (MT-Bench ~6.64), but dropped from 2024-25 default SFT mixes (absent from SmolTalk and Tulu 3).

https://ar5iv.labs.arxiv.org/html/2310.16944 recorded 2026-07-04
Zephyr: dSFT on UltraChat reaches MT-Bench 6.64
https://huggingface.co/datasets/allenai/tulu-3-sft-mixture recorded 2026-07-04
UltraChat-200k is not used as an SFT source in Tulu 3

Unchanged since 2026-07-04 (last edited, not re-checked)