Add files using upload-large-folder tool

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.0.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 645, "1": 0, "2": 0, "3": 0}, "net_accuracy": 0.21371769383697814}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.1.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 69, "1": 10, "2": 12, "3": 706}, "net_accuracy": 0.26408217362491715}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.10.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 445, "1": 107, "2": 0, "3": 103}, "net_accuracy": 0.2170311464546057}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.11.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 19, "1": 702, "2": 0, "3": 2}, "net_accuracy": 0.23956262425447317}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.11.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 522, "1": 39, "2": 59, "3": 33}, "net_accuracy": 0.21636845593108017}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.12.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 4, "1": 706, "2": 1, "3": 14}, "net_accuracy": 0.24022531477799866}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.12.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 616, "1": 23, "2": 5, "3": 6}, "net_accuracy": 0.21537442014579192}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.13.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 552, "1": 2, "2": 120, "3": 0}, "net_accuracy": 0.22332670642809807}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.13.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 575, "1": 3, "2": 72, "3": 12}, "net_accuracy": 0.219350563286945}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.14.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 719, "2": 0, "3": 0}, "net_accuracy": 0.23823724320742212}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.14.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 360, "1": 46, "2": 182, "3": 127}, "net_accuracy": 0.2369118621603711}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.15.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 121, "1": 388, "2": 77, "3": 171}, "net_accuracy": 0.2508283631544069}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.16.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 6, "1": 701, "2": 3, "3": 1}, "net_accuracy": 0.2355864811133201}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.16.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 459, "1": 111, "2": 57, "3": 74}, "net_accuracy": 0.2322730284956925}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.17.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 341, "2": 376, "3": 77}, "net_accuracy": 0.26308813783962887}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.17.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 384, "1": 95, "2": 181, "3": 52}, "net_accuracy": 0.23591782637508282}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.18.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 448, "1": 22, "2": 203, "3": 30}, "net_accuracy": 0.23293571901921803}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.19.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 165, "1": 115, "2": 275, "3": 273}, "net_accuracy": 0.27435387673956263}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.19.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 642, "1": 0, "2": 7, "3": 1}, "net_accuracy": 0.21537442014579192}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.2.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 18, "1": 550, "2": 174, "3": 0}, "net_accuracy": 0.24585818422796554}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.2.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 0, "2": 0, "3": 871}, "net_accuracy": 0.2886017229953612}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.20.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 36, "2": 458, "3": 273}, "net_accuracy": 0.2541418157720345}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.20.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 623, "1": 0, "2": 42, "3": 11}, "net_accuracy": 0.2239893969516236}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.21.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 12, "1": 14, "2": 414, "3": 369}, "net_accuracy": 0.2680583167660702}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.21.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 618, "1": 3, "2": 35, "3": 11}, "net_accuracy": 0.22100728959575877}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.22.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 562, "1": 0, "2": 98, "3": 18}, "net_accuracy": 0.22465208747514911}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.23.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 425, "2": 340, "3": 2}, "net_accuracy": 0.2541418157720345}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.23.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 227, "1": 19, "2": 439, "3": 140}, "net_accuracy": 0.27335984095427435}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.24.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 645, "1": 1, "2": 0, "3": 0}, "net_accuracy": 0.2140490390987409}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.24.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 413, "1": 14, "2": 274, "3": 52}, "net_accuracy": 0.24950298210735586}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.25.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 22, "1": 65, "2": 443, "3": 211}, "net_accuracy": 0.24552683896620278}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.25.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 336, "1": 21, "2": 267, "3": 178}, "net_accuracy": 0.26573889993373095}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.26.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 1, "1": 0, "2": 2, "3": 855}, "net_accuracy": 0.28429423459244535}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.26.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 76, "1": 47, "2": 112, "3": 646}, "net_accuracy": 0.29191517561298874}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.27.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 117, "1": 454, "2": 105, "3": 0}, "net_accuracy": 0.2239893969516236}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.3.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 11, "2": 224, "3": 587}, "net_accuracy": 0.27236580516898606}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.3.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 0, "2": 0, "3": 871}, "net_accuracy": 0.2886017229953612}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.4.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 540, "1": 0, "2": 102, "3": 30}, "net_accuracy": 0.22266401590457258}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.5.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 29, "1": 699, "2": 0, "3": 0}, "net_accuracy": 0.24121935056328694}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.6.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 276, "1": 234, "2": 216, "3": 0}, "net_accuracy": 0.24055666003976142}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.6.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 129, "1": 12, "2": 34, "3": 622}, "net_accuracy": 0.26408217362491715}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.7.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 37, "1": 664, "2": 8, "3": 0}, "net_accuracy": 0.23492379058979457}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.7.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 430, "1": 70, "2": 25, "3": 152}, "net_accuracy": 0.22432074221338635}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.8.hook_mlp_out.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 15, "1": 48, "2": 120, "3": 629}, "net_accuracy": 0.2690523525513585}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.8.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 427, "1": 204, "2": 7, "3": 25}, "net_accuracy": 0.21968190854870775}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.9.hook_resid_post.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 338, "1": 64, "2": 150, "3": 154}, "net_accuracy": 0.23392975480450628}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_ln_final.hook_normalized.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 197, "1": 203, "2": 283, "3": 121}, "net_accuracy": 0.2664015904572565}

mmlu_STEM/few_shot_2/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_output.json

@@ -0,0 +1 @@
+{"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 197, "1": 203, "2": 283, "3": 121}, "net_accuracy": 0.2664015904572565}