datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
pretext-ui-harbor-runs-v0
Pretext UI Harbor Runs
Harbor task-generation and solve-run corpus for the @chenglou/pretext UI task family. The dataset contains flat training indexes plus raw redacted Harbor artifacts.
Contents
data/train/attempts.jsonl: one row per candidate attempt, with model bucket, reward, Gemini score, prompt, and raw artifact pointers.
data/train/tasks.jsonl: materialized task identity/hash index.
data/train/sft_conversations.jsonl: OpenAI-style user/assistant conversation rows… See the full description on the dataset page: https://huggingface.co/datasets/RLAIF/pretext-ui-harbor-runs-v0.pretextucount_pretext