ultrastar111/

pusht_96_norm4_cot_chunk_k10_20260622_perseg

PushT (96px, norm4, JPEG q90; coverage task, no hard split — in-dist claims only) — action-conditioned visual world-model SFT data (CoT self-rollout) for the BAGEL-7B-MoT feedback-interval study.

Format: gzipped JSONL shards under training/, 1 row = 1 packed episode. CoT rows: per-segment layout — per-step imagined frame (MSE target) + committed action chunk, with a loss-0 "Action executed." + real frame re-grounding… See the full description on the dataset page: https://huggingface.co/datasets/ultrastar111/pusht_96_norm4_cot_chunk_k10_20260622_perseg.