{"activation": "topk", "expansion_factor": 2, "normalize_decoder": true, "num_latents": 0, "k": 32, "multi_topk": false, "skip_connection": false, "transcode": false, "d_in": 1152} |
{"activation": "topk", "expansion_factor": 2, "normalize_decoder": true, "num_latents": 0, "k": 32, "multi_topk": false, "skip_connection": false, "transcode": false, "d_in": 1152} |