diff --git a/attn_out_all/layer_0_width_16k_l0_big/config.json b/attn_out_all/layer_0_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..56f8c9f5918a8a1b74bc32654f7ea5007ca21478 --- /dev/null +++ b/attn_out_all/layer_0_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.0.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.0.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 60, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_0_width_16k_l0_small/config.json b/attn_out_all/layer_0_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..0e023bb185cf18dc9fe8a88e18ecd6ce1f4b916c --- /dev/null +++ b/attn_out_all/layer_0_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.0.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.0.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 10, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_0_width_262k_l0_big/config.json b/attn_out_all/layer_0_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..35fa7c721e5edbb823df4fe87b6123db324a1ea0 --- /dev/null +++ b/attn_out_all/layer_0_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.0.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.0.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 60, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_0_width_262k_l0_small/config.json b/attn_out_all/layer_0_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..00845ad064492a16cda33296a5c650c98a8ef9fd --- /dev/null +++ b/attn_out_all/layer_0_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.0.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.0.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 10, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_10_width_16k_l0_big/config.json b/attn_out_all/layer_10_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..3cc8e935bec67ff670f00df141044597c3ae1bd8 --- /dev/null +++ b/attn_out_all/layer_10_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.10.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.10.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_10_width_262k_l0_big/config.json b/attn_out_all/layer_10_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..aebaace57b98e9b7905a0e723b79f35b3d8da033 --- /dev/null +++ b/attn_out_all/layer_10_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.10.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.10.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_11_width_16k_l0_big/config.json b/attn_out_all/layer_11_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..b57c31f21276af1fb2d9a250db73d0aae9e6d097 --- /dev/null +++ b/attn_out_all/layer_11_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.11.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.11.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_11_width_16k_l0_small/config.json b/attn_out_all/layer_11_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..349a63715e451a1e2815b6505425207bf6abd4e1 --- /dev/null +++ b/attn_out_all/layer_11_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.11.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.11.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_11_width_262k_l0_small/config.json b/attn_out_all/layer_11_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c7b8a44c9c79fecbb8efa0df89728c5063d208a8 --- /dev/null +++ b/attn_out_all/layer_11_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.11.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.11.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_13_width_16k_l0_big/config.json b/attn_out_all/layer_13_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..a1352ac1adda815076bd44546281e1dc4f18b238 --- /dev/null +++ b/attn_out_all/layer_13_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.13.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.13.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_13_width_16k_l0_small/config.json b/attn_out_all/layer_13_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..a7c1d3eda57fa81f6d5ceb42e04e37bb54a95c5d --- /dev/null +++ b/attn_out_all/layer_13_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.13.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.13.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_14_width_16k_l0_big/config.json b/attn_out_all/layer_14_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..eeb37ded0c3252d1650383405e2cbb5eada28ac0 --- /dev/null +++ b/attn_out_all/layer_14_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.14.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.14.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_14_width_262k_l0_small/config.json b/attn_out_all/layer_14_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..11cb6cb7a51a1a230a40b5588e39f0a3240d59d0 --- /dev/null +++ b/attn_out_all/layer_14_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.14.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.14.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_15_width_16k_l0_big/config.json b/attn_out_all/layer_15_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..2c9f88ab44cc3a8cd8866d5ccbb65b7a56a38e9c --- /dev/null +++ b/attn_out_all/layer_15_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.15.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.15.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_15_width_262k_l0_big/config.json b/attn_out_all/layer_15_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..d88640d42636e1feb655f5f09b9047b9479bd9d9 --- /dev/null +++ b/attn_out_all/layer_15_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.15.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.15.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_16_width_262k_l0_small/config.json b/attn_out_all/layer_16_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..3c2e2f1ea4027718c64ac052f37c44ab11e246f9 --- /dev/null +++ b/attn_out_all/layer_16_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.16.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.16.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_17_width_16k_l0_big/config.json b/attn_out_all/layer_17_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..bf66d965d64d227b1561bae16b482e528724edb0 --- /dev/null +++ b/attn_out_all/layer_17_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.17.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.17.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_17_width_262k_l0_big/config.json b/attn_out_all/layer_17_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..f68b07b23d174ef19d639be50ea644d6c4af433d --- /dev/null +++ b/attn_out_all/layer_17_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.17.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.17.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_17_width_262k_l0_small/config.json b/attn_out_all/layer_17_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..33f0b4c2c95bef7aa68b329278cf98614e576796 --- /dev/null +++ b/attn_out_all/layer_17_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.17.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.17.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_18_width_16k_l0_big/config.json b/attn_out_all/layer_18_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..3cebf6a00dc88385298f845f0369d3c0a3e1242e --- /dev/null +++ b/attn_out_all/layer_18_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.18.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.18.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_19_width_16k_l0_big/config.json b/attn_out_all/layer_19_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..4308d712bfcbc09dc83573b70d4f250f032e5f35 --- /dev/null +++ b/attn_out_all/layer_19_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.19.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.19.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_19_width_16k_l0_small/config.json b/attn_out_all/layer_19_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..d3b9986eafb46eef8d11446ff71b04d3fef1c607 --- /dev/null +++ b/attn_out_all/layer_19_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.19.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.19.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_19_width_262k_l0_small/config.json b/attn_out_all/layer_19_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..196ad8560eb16df3e9c374a5d938bfc82c13400f --- /dev/null +++ b/attn_out_all/layer_19_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.19.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.19.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_20_width_16k_l0_small/config.json b/attn_out_all/layer_20_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..58fdd445023d07f3602241319c586f5bf2e595da --- /dev/null +++ b/attn_out_all/layer_20_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.20.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.20.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_20_width_262k_l0_big/config.json b/attn_out_all/layer_20_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..5d02a802a1c8e640ed35b84a47ec6fdac03b87a5 --- /dev/null +++ b/attn_out_all/layer_20_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.20.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.20.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_20_width_262k_l0_small/config.json b/attn_out_all/layer_20_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..d22b91058053206c3af41d5f11b71684c7603f4c --- /dev/null +++ b/attn_out_all/layer_20_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.20.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.20.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_21_width_16k_l0_small/config.json b/attn_out_all/layer_21_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..822160b11b36644de6ff01cdc875617224763948 --- /dev/null +++ b/attn_out_all/layer_21_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.21.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.21.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_21_width_262k_l0_small/config.json b/attn_out_all/layer_21_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..6807a8ef9218e949e096b9940e76629eb50c594c --- /dev/null +++ b/attn_out_all/layer_21_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.21.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.21.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_22_width_16k_l0_big/config.json b/attn_out_all/layer_22_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..10980605f07648c5100c465696b3e4c750b7fe73 --- /dev/null +++ b/attn_out_all/layer_22_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.22.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.22.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_22_width_16k_l0_small/config.json b/attn_out_all/layer_22_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..f75d9d2725215cbb939837e27cf57bbad2aa26b7 --- /dev/null +++ b/attn_out_all/layer_22_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.22.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.22.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_22_width_262k_l0_big/config.json b/attn_out_all/layer_22_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..03514ca5163ea968509187177aeb0fea8a0bbbf4 --- /dev/null +++ b/attn_out_all/layer_22_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.22.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.22.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_23_width_16k_l0_big/config.json b/attn_out_all/layer_23_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..092a2a665924b5b37ded222576647d4f788430f3 --- /dev/null +++ b/attn_out_all/layer_23_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.23.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.23.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_23_width_262k_l0_big/config.json b/attn_out_all/layer_23_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..feaa7a5e6af48ffd3391f973b5928fbfbba0ebb6 --- /dev/null +++ b/attn_out_all/layer_23_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.23.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.23.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_23_width_262k_l0_small/config.json b/attn_out_all/layer_23_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..4f67bb753f17a9ece0c2a719a2001795671490a8 --- /dev/null +++ b/attn_out_all/layer_23_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.23.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.23.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_24_width_262k_l0_big/config.json b/attn_out_all/layer_24_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..f686ccf724ae8fea7e7acc1139923d20b08d901e --- /dev/null +++ b/attn_out_all/layer_24_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.24.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.24.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_25_width_262k_l0_big/config.json b/attn_out_all/layer_25_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..1ccffc1e73e0bd4da79e4ffe49fd7cbdfb20cefa --- /dev/null +++ b/attn_out_all/layer_25_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.25.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.25.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_2_width_16k_l0_small/config.json b/attn_out_all/layer_2_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..a499a849439159bc1ea0dad2bc03969687ec2ae2 --- /dev/null +++ b/attn_out_all/layer_2_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.2.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.2.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 12, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_2_width_262k_l0_big/config.json b/attn_out_all/layer_2_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..d4399f73684ea5ac4d8130293495d8c550429ebf --- /dev/null +++ b/attn_out_all/layer_2_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.2.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.2.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 73, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_2_width_262k_l0_small/config.json b/attn_out_all/layer_2_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..64cbb2941e82fe18fe884da3111477ca187caa6f --- /dev/null +++ b/attn_out_all/layer_2_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.2.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.2.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 12, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_3_width_16k_l0_big/config.json b/attn_out_all/layer_3_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..261e4488b11ab3e703a03739bc81a06114c616f7 --- /dev/null +++ b/attn_out_all/layer_3_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.3.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.3.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 80, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_3_width_16k_l0_small/config.json b/attn_out_all/layer_3_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..154d4d2f8f5112124a16b3bae391c94a2b6bf712 --- /dev/null +++ b/attn_out_all/layer_3_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.3.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.3.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 13, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_3_width_262k_l0_small/config.json b/attn_out_all/layer_3_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..f01bb890ed113b5dfc5af56dcd9dabbce9f09ab3 --- /dev/null +++ b/attn_out_all/layer_3_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.3.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.3.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 13, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_4_width_16k_l0_big/config.json b/attn_out_all/layer_4_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..d9a9cb0b25036f5ee04b1e04e23e793c3a2b20ef --- /dev/null +++ b/attn_out_all/layer_4_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.4.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.4.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 87, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_4_width_262k_l0_big/config.json b/attn_out_all/layer_4_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..bfa85ac2d180c6fa0149234e384a3af6eea6527f --- /dev/null +++ b/attn_out_all/layer_4_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.4.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.4.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 87, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_5_width_16k_l0_big/config.json b/attn_out_all/layer_5_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..e2ddd9b27b98e08b8103e48af555d52d7e86bf22 --- /dev/null +++ b/attn_out_all/layer_5_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.5.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.5.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 94, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_5_width_16k_l0_small/config.json b/attn_out_all/layer_5_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..84323f84cc0e37c8c51188c4e06abcc22d687a27 --- /dev/null +++ b/attn_out_all/layer_5_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.5.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.5.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 15, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_5_width_262k_l0_big/config.json b/attn_out_all/layer_5_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..83be7b77fcf41f7a080f9ea9ded8f1f8c0b4f5fb --- /dev/null +++ b/attn_out_all/layer_5_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.5.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.5.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 94, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_6_width_16k_l0_small/config.json b/attn_out_all/layer_6_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..d939669579e583f8a4825bebe6d0a08cc80fb7e8 --- /dev/null +++ b/attn_out_all/layer_6_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.6.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.6.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 16, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_7_width_16k_l0_big/config.json b/attn_out_all/layer_7_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..fffa3493fc915af18ef9c3dd6a9616d98233f84f --- /dev/null +++ b/attn_out_all/layer_7_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.7.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.7.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 108, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_7_width_262k_l0_big/config.json b/attn_out_all/layer_7_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..8099591aa12e769b4af5d1e357694ba2177555cc --- /dev/null +++ b/attn_out_all/layer_7_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.7.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.7.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 108, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_7_width_262k_l0_small/config.json b/attn_out_all/layer_7_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..2c38831e0884b198cfdb45adba8e5193284be5c8 --- /dev/null +++ b/attn_out_all/layer_7_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.7.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.7.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 18, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_8_width_16k_l0_big/config.json b/attn_out_all/layer_8_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..7391986f113659e48a2bb0d70a167843bf4bffd8 --- /dev/null +++ b/attn_out_all/layer_8_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.8.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.8.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 115, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_8_width_262k_l0_big/config.json b/attn_out_all/layer_8_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..f9042c46bad695b4fdd1683d8716afbc8e89db97 --- /dev/null +++ b/attn_out_all/layer_8_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.8.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.8.self_attn.o_proj.input", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 115, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/attn_out_all/layer_9_width_16k_l0_small/config.json b/attn_out_all/layer_9_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..6d031d3e581ad3685c52154d0a62213e02dbd478 --- /dev/null +++ b/attn_out_all/layer_9_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.9.self_attn.o_proj.input", + "hf_hook_point_out": "model.layers.9.self_attn.o_proj.input", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "sae" +} \ No newline at end of file diff --git a/transcoder_all/layer_0_width_16k_l0_small_affine/config.json b/transcoder_all/layer_0_width_16k_l0_small_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..c50546d5987b340e732c171d8bf75ccd74848437 --- /dev/null +++ b/transcoder_all/layer_0_width_16k_l0_small_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.0.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.0.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 10, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_12_width_262k_l0_small/config.json b/transcoder_all/layer_12_width_262k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..b4f15a2a8aa2f83d1bf1f75aaf883f3c7daa695d --- /dev/null +++ b/transcoder_all/layer_12_width_262k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.12.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.12.post_feedforward_layernorm.output", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_14_width_262k_l0_big/config.json b/transcoder_all/layer_14_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..495d38cf451379e919a21a59f399ed6b9130bfa0 --- /dev/null +++ b/transcoder_all/layer_14_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.14.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.14.post_feedforward_layernorm.output", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_14_width_262k_l0_small_affine/config.json b/transcoder_all/layer_14_width_262k_l0_small_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..681a166387d169d132fe626036b8563e0c0cc728 --- /dev/null +++ b/transcoder_all/layer_14_width_262k_l0_small_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.14.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.14.post_feedforward_layernorm.output", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_16_width_16k_l0_big_affine/config.json b/transcoder_all/layer_16_width_16k_l0_big_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..ce5d7aa26b7048140d89c476467c1ea26bf0bdab --- /dev/null +++ b/transcoder_all/layer_16_width_16k_l0_big_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.16.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.16.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_16_width_16k_l0_small/config.json b/transcoder_all/layer_16_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..7526e9acb586546e422c857848e8f4327b9d416a --- /dev/null +++ b/transcoder_all/layer_16_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.16.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.16.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_16_width_16k_l0_small_affine/config.json b/transcoder_all/layer_16_width_16k_l0_small_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..bcdcaf97e971475ec4283ff531fa0c7930c5fab1 --- /dev/null +++ b/transcoder_all/layer_16_width_16k_l0_small_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.16.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.16.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_16_width_262k_l0_big/config.json b/transcoder_all/layer_16_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..2a616d148977999a917bd94ecd3f44a40a18ac03 --- /dev/null +++ b/transcoder_all/layer_16_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.16.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.16.post_feedforward_layernorm.output", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_18_width_16k_l0_small/config.json b/transcoder_all/layer_18_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..0c5c215e37ac62954c6f68bcaaf4539e798d384d --- /dev/null +++ b/transcoder_all/layer_18_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.18.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.18.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_19_width_262k_l0_big_affine/config.json b/transcoder_all/layer_19_width_262k_l0_big_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..17e038e557ab890a38f639112cb3e3aa3f3da5bb --- /dev/null +++ b/transcoder_all/layer_19_width_262k_l0_big_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.19.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.19.post_feedforward_layernorm.output", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_1_width_16k_l0_small_affine/config.json b/transcoder_all/layer_1_width_16k_l0_small_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..ca69495781aadc0416adf6df223771cc217c00d3 --- /dev/null +++ b/transcoder_all/layer_1_width_16k_l0_small_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.1.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.1.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 11, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_1_width_262k_l0_big/config.json b/transcoder_all/layer_1_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..d0457826248b96861b4b1152a2b10cfa13ce9a5c --- /dev/null +++ b/transcoder_all/layer_1_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.1.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.1.post_feedforward_layernorm.output", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 66, + "affine_connection": false, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_20_width_16k_l0_big/config.json b/transcoder_all/layer_20_width_16k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..58d9770bb2973455306f25465f75f1af485e820e --- /dev/null +++ b/transcoder_all/layer_20_width_16k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.20.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.20.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_21_width_16k_l0_small_affine/config.json b/transcoder_all/layer_21_width_16k_l0_small_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..b8ef178f32a08ca9486132cf08068bcafc75b8fd --- /dev/null +++ b/transcoder_all/layer_21_width_16k_l0_small_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.21.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.21.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_21_width_262k_l0_big/config.json b/transcoder_all/layer_21_width_262k_l0_big/config.json new file mode 100644 index 0000000000000000000000000000000000000000..f9e6c311797391c18ed612698126946349cf3f21 --- /dev/null +++ b/transcoder_all/layer_21_width_262k_l0_big/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.21.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.21.post_feedforward_layernorm.output", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": false, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_23_width_16k_l0_small/config.json b/transcoder_all/layer_23_width_16k_l0_small/config.json new file mode 100644 index 0000000000000000000000000000000000000000..8c9f9767deb37c05316a248f589f13cf0815e169 --- /dev/null +++ b/transcoder_all/layer_23_width_16k_l0_small/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.23.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.23.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": false, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_25_width_16k_l0_small_affine/config.json b/transcoder_all/layer_25_width_16k_l0_small_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..89805053633ecd1fa2c02cf8e1c732240949ebfa --- /dev/null +++ b/transcoder_all/layer_25_width_16k_l0_small_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.25.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.25.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 20, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_3_width_16k_l0_big_affine/config.json b/transcoder_all/layer_3_width_16k_l0_big_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..2687e5f2ee8dabca2b7fb6d86fc3ddaacc274604 --- /dev/null +++ b/transcoder_all/layer_3_width_16k_l0_big_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.3.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.3.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 80, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_4_width_262k_l0_big_affine/config.json b/transcoder_all/layer_4_width_262k_l0_big_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..ebf225a951874431a3849b0497be68b0660f4175 --- /dev/null +++ b/transcoder_all/layer_4_width_262k_l0_big_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.4.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.4.post_feedforward_layernorm.output", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 87, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_6_width_16k_l0_small_affine/config.json b/transcoder_all/layer_6_width_16k_l0_small_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..a1f930a4f904b0e2dc9e2aa9e20b674c62e259b4 --- /dev/null +++ b/transcoder_all/layer_6_width_16k_l0_small_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.6.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.6.post_feedforward_layernorm.output", + "width": 16384, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 16, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file diff --git a/transcoder_all/layer_9_width_262k_l0_big_affine/config.json b/transcoder_all/layer_9_width_262k_l0_big_affine/config.json new file mode 100644 index 0000000000000000000000000000000000000000..a8c3db8f352b555f88178234f1b0fe3a962506c9 --- /dev/null +++ b/transcoder_all/layer_9_width_262k_l0_big_affine/config.json @@ -0,0 +1,10 @@ +{ + "hf_hook_point_in": "model.layers.9.pre_feedforward_layernorm.output", + "hf_hook_point_out": "model.layers.9.post_feedforward_layernorm.output", + "width": 262144, + "model_name": "google/gemma-3-1b-it", + "architecture": "jump_relu", + "l0": 120, + "affine_connection": true, + "type": "transcoder" +} \ No newline at end of file