apart
/

saes-llama3.2-1b-base-smollm-1m

Model card Files Files and versions

Model Card for Model ID

Trained SAE's using EleutherAI's sparsity package on meta-llama/Llama-3.2-1B

See saes/wandb_metrics.json for various stats on fvu and dead feature pct per layer.

Downloads last month: 32

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for apart/saes-llama3.2-1b-base-smollm-1m

Base model

meta-llama/Llama-3.2-1B

Finetuned

(734)

this model

Dataset used to train apart/saes-llama3.2-1b-base-smollm-1m