Kalso42/WorldModelForMaze
WorldModelForMaze Code, datasets, and trained checkpoints for studying world-model representations in maze navigation, based on a modified NanoGPT. Contents *.py — training, testing, probing, and visualization scripts (see readme.md). model/ — architectures: transformer, transformer-rope, transformer-nextlat, mamba, mamba2, gated-deltanet, gru. data/maze/100/ — tokenized maze datasets for Tasks A/C/E/H/I (RWs paths, 100 nodes). out/ — final (10000-iter)… See the full description on the dataset page: https://huggingface.co/datasets/Kalso42/WorldModelForMaze.
Upload out/plot/probe_target_dist_H1_10M_12_12_576_24_576.npz with huggingface_hub
Upload out/plot/probe_target_dist_H1_10M_12_12_576_24_576.png with huggingface_hub
Upload out/plot/probe_target_dist_I1_10M_6_6_384_12_384.npz with huggingface_hub
Upload out/plot/probe_target_dist_I1_10M_6_6_384_12_384.png with huggingface_hub
Upload out/plot/probe_target_dist_H1_10M_12_12_576_24_576.png with huggingface_hub
Upload out/plot/probe_target_dist_I1_10M_6_6_384_12_384.png with huggingface_hub
Upload out/plot/probe_target_dist_C1_10M_6_6_384_12_384.npz with huggingface_hub
Upload out/plot/probe_target_dist_C1_10M_6_6_384_12_384.png with huggingface_hub
Upload out/plot/probe_target_dist_E1_500K_3_1_256_6_256.npz with huggingface_hub
Upload out/plot/probe_target_dist_E1_500K_3_1_256_6_256.png with huggingface_hub
Upload out/plot/probe_target_dist_A1_500K_3_1_256_6_256.npz with huggingface_hub
Upload out/plot/probe_target_dist_A1_500K_3_1_256_6_256.png with huggingface_hub
Upload folder using huggingface_hub (part 3)
Upload folder using huggingface_hub (part 2)
Upload folder using huggingface_hub
Upload README.md with huggingface_hub
Upload folder using huggingface_hub (part 4)
Upload folder using huggingface_hub (part 3)
Upload folder using huggingface_hub (part 2)
Upload folder using huggingface_hub
initial commit