Instructions to use apart/llama3.2_1b_instruct_saes_vader with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use apart/llama3.2_1b_instruct_saes_vader with Transformers:
# pip install -U transformers accelerate # Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("apart/llama3.2_1b_instruct_saes_vader", device_map="auto") - Notebooks
- Google Colab
- Kaggle
File size: 775 Bytes
fb05c99 | 1 | {"sae": {"expansion_factor": 4, "normalize_decoder": true, "num_latents": 0, "k": 128, "multi_topk": false, "skip_connection": false}, "batch_size": 8, "grad_acc_steps": 1, "micro_acc_steps": 1, "lr": 0.0005, "lr_warmup_steps": 1000, "auxk_alpha": 0.0, "dead_feature_threshold": 10000000, "hookpoints": ["model.layers.2", "model.layers.3", "model.layers.4", "model.layers.5", "model.layers.6", "model.layers.7", "model.layers.8", "model.layers.9", "model.layers.10", "model.layers.11", "model.layers.12", "model.layers.13", "model.layers.14"], "init_seeds": [0], "layers": [], "layer_stride": 1, "transcode": false, "distribute_modules": false, "save_every": 100000, "log_to_wandb": true, "run_name": "saes_meta-llama/Llama-3.2-1B-Instruct/k=128", "wandb_log_frequency": 100} |