0

UltraChat

Fresh

Tsinghua / OpenBMB's large-scale multi-turn dialog dataset generated by two LLMs talking to each other across structured topic taxonomies.

Type
SFT Dataset
Publisher
OpenBMB
Runtime
hf_parquet
License
MIT
Size
1.5M dialogs (~200k in the popular "200k" filter)
Published
May 2026

Cite

Notes

Only stored in your browser.

Lift evidence

2
EvalTools known to liftSource paper
MT-BenchUltraChat-
AlpacaEvalUltraChat-

Models

Notable models trained on it

UltraLM-13BZephyr-7B (SFT phase)many Hugging Face H4 derivative models

Papers

1

Contributors

3