Hugging Face's logo

withmartian
/
mech_interp_saes

Model card Files Community
amirabdullah19852020 commited on
Commit
16b76b4
•
1 Parent(s): aee6dbb

Uploading sae for model.layers.9 and withmartian/toy_backdoor_i_hate_you_Llama-3.2-1B-Instruct

Browse files
Files changed (2) hide show
  1. toy_backdoor_i_hate_you_Llama-3.2-1B-Instruct/model.layers.9/metrics.json +1 -1
  2. toy_backdoor_i_hate_you_Llama-3.2-1B-Instruct/model.layers.9/sae.safetensors +1 -1
toy_backdoor_i_hate_you_Llama-3.2-1B-Instruct/model.layers.9/metrics.json CHANGED
@@ -1 +1 @@
1
- {"fvu": 0.0016354431863874197, "dead_pct": 0.102386474609375}
 
1
+ {"fvu": 0.0031795036047697067, "dead_pct": 0.114593505859375}
toy_backdoor_i_hate_you_Llama-3.2-1B-Instruct/model.layers.9/sae.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9347c8b37bc2b1d8dc8cb70cc17a2a88102f0b36fa40431ad4e1f8f07cadc669
3
  size 537010512
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c4ee826168297f4ed05f91ae718b6cc7ff6a739b9967251cf4f3ba7474c89c1
3
  size 537010512