Always-on ordinary one-branch counterfactual training data used by the Table-2 branch baseline. layout: Train-only. Gzipped-JSONL shards under training/; each row is one packed SFT sample with base64-JPEG frames inline. images are base64-encoded JPEG frames stored inline in each JSONL row.

Pairs with the matching sokoban checkpoint(s) under the companion model org; CoT and non-CoT… See the full description on the dataset page: https://huggingface.co/datasets/novastar111/sokoban_adaptive_ordinary_branch_train.