datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
stargate_s04e01_100topkdiverse_text2vid
LLaVA-CoT-30k-base64-in-jsonlfullstackarena-table-study-v1
FullStackArena three-site table study (private working dataset)
Protocol fullstackarena-three-site-table-study-v1 (status qualified): five exact OpenRouter models × three cells × the signed Arenagram (124), RideApp (120) and Auction (100) task packs = 45 arms, 5,160 scored attempts, max_steps=30, five workers on five isolated lanes per batch.
Cells: table1_no_time_capped, table1_no_time_uncapped (no automatic time; the explicit get_website_time() tool stays available)… See the full description on the dataset page: https://huggingface.co/datasets/GroupieSteven/fullstackarena-table-study-v1.
