SAEs
Collection
Sparse Autoencoders trained by qresearch
•
2 items
•
Updated
Model Information
A SAE (Sparse Autoencoder) for deepseek-ai/DeepSeek-R1-Distill-Llama-8B.
It is trained specifically on layer 19 of DeepSeek-R1-Distill-Llama-8B and achieves a final L0 of 93 during training.
This model is used to decompose Llama's activations into interpretable features.
The SAE weights are released under Apache, however DeepSeek-R1-Distill-Llama-8B is to be used under Meta's Llama 3 License.
How to use
A Jupyter Notebook is provided to test the model
Training
Our SAE was trained using LMSYS-Chat-1M dataset.
Acknowledgements
This release wouldn't have been possible without the work of Goodfire and Anthropic
A huge thank goes to runpod, who generously sponsored the compute for this run!
.x+=:.
z` ^% .uef^"
.u . . <k .u . :d88E
.u@u .d88B :@8c .u .@8Ned8" .u u .d88B :@8c . `888E
.zWF8888bx ="8888f8888r ud8888. .@^%8888" ud8888. us888u. ="8888f8888r .udR88N 888E .z8k
.888 9888 4888>'88" :888'8888. x88: `)8b. :888'8888. .@88 "8888" 4888>'88" <888'888k 888E~?888L
I888 9888 4888> ' d888 '88%" 8888N=*8888 d888 '88%" 9888 9888 4888> ' 9888 'Y" 888E 888E
I888 9888 4888> 8888.+" %8" R88 8888.+" 9888 9888 4888> 9888 888E 888E
I888 9888 .d888L .+ 8888L @8Wou 9% 8888L 9888 9888 .d888L .+ 9888 888E 888E
`888Nx?888 ^"8888*" '8888c. .+ .888888P` '8888c. .+ 9888 9888 ^"8888*" ?8888u../ 888E 888E
"88" '888 "Y" "88888% ` ^"F "88888% "888*""888" "Y" "8888P' m888N= 888>
88E "YP' "YP' ^Y" ^Y' "P' `Y" 888
98> J88"
'8 @%
` :"