datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
constellation-bench
ConstellationBench: Behavioral AI Evaluation Across 22 LLM Models
Alignment made frontier models worse at being someone. We built an open benchmark that shows it.
A free Qwen model scores 0.617 on persona fidelity. Anthropic's brand-new Opus 4.7 scores 0.538. Google's open-weight Gemma-4 beats it at 38x less cost. This pattern held across every architecture we tested -- dense transformers, MoE, Mamba-Transformer hybrids, and linear attention. 22 models. 22,200+ LLM calls. $115… See the full description on the dataset page: https://huggingface.co/datasets/AirlockLabs/constellation-bench.pega-constellation-dx-sft
Pega Constellation DX Components — Training Dataset
This dataset teaches a code model how to write Pega Constellation DX components — the custom React/TypeScript components that extend the Pega Platform UI.
It comes from the open-source constellation-ui-gallery repo, which Pega itself maintains as a reference for DX component authors. We pinned a specific commit so the dataset is reproducible.
What is in it
55 Pega DX components, each broken into 9 different… See the full description on the dataset page: https://huggingface.co/datasets/WeekendNoobs/pega-constellation-dx-sft.