Hugging Face's logo

Datasets:
Exploration-Lab
/
dim-discovery-archive

ArXiv:
Dataset card Files
xet
 Community
Divyanshu-Bhatt commited on
Commit
e1e2192
·
verified ·
1 Parent(s): d23279b

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. mmlu_STEM/few_shot_1/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.19.hook_resid_post.json +1 -0
  2. mmlu_STEM/few_shot_1/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.21.hook_resid_post.json +1 -0
  3. mmlu_STEM/few_shot_1/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.23.hook_mlp_out.json +1 -0
  4. mmlu_STEM/few_shot_1/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.24.hook_resid_post.json +1 -0
  5. mmlu_STEM/few_shot_1/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_ln_final.hook_normalized.json +1 -0
  6. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.0.hook_mlp_out.json +1 -0
  7. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.0.hook_resid_post.json +1 -0
  8. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.1.hook_mlp_out.json +1 -0
  9. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.1.hook_resid_post.json +1 -0
  10. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.10.hook_mlp_out.json +1 -0
  11. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.10.hook_resid_post.json +1 -0
  12. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.11.hook_mlp_out.json +1 -0
  13. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.11.hook_resid_post.json +1 -0
  14. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.2.hook_mlp_out.json +1 -0
  15. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.2.hook_resid_post.json +1 -0
  16. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.3.hook_mlp_out.json +1 -0
  17. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.3.hook_resid_post.json +1 -0
  18. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.4.hook_mlp_out.json +1 -0
  19. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.5.hook_mlp_out.json +1 -0
  20. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.5.hook_resid_post.json +1 -0
  21. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.6.hook_mlp_out.json +1 -0
  22. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.6.hook_resid_post.json +1 -0
  23. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.7.hook_mlp_out.json +1 -0
  24. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.7.hook_resid_post.json +1 -0
  25. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.8.hook_mlp_out.json +1 -0
  26. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.8.hook_resid_post.json +1 -0
  27. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.9.hook_mlp_out.json +1 -0
  28. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.9.hook_resid_post.json +1 -0
  29. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_ln_final.hook_normalized.json +1 -0
  30. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_output.json +1 -0
  31. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.0.hook_mlp_out.csv +0 -0
  32. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.0.hook_resid_post.csv +0 -0
  33. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.1.hook_mlp_out.csv +0 -0
  34. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.1.hook_resid_post.csv +0 -0
  35. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.10.hook_mlp_out.csv +0 -0
  36. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.10.hook_resid_post.csv +0 -0
  37. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.11.hook_mlp_out.csv +0 -0
  38. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.11.hook_resid_post.csv +0 -0
  39. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.2.hook_mlp_out.csv +0 -0
  40. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.2.hook_resid_post.csv +0 -0
  41. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.3.hook_mlp_out.csv +0 -0
  42. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.3.hook_resid_post.csv +0 -0
  43. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.4.hook_mlp_out.csv +0 -0
  44. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.4.hook_resid_post.csv +0 -0
  45. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.5.hook_mlp_out.csv +0 -0
  46. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.5.hook_resid_post.csv +0 -0
  47. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.6.hook_mlp_out.csv +0 -0
  48. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.6.hook_resid_post.csv +0 -0
  49. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.7.hook_mlp_out.csv +0 -0
  50. mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.7.hook_resid_post.csv +0 -0
mmlu_STEM/few_shot_1/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.19.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 639, "1": 1, "2": 9, "3": 1}, "net_accuracy": 0.21537442014579192}
mmlu_STEM/few_shot_1/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.21.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 604, "1": 12, "2": 46, "3": 10}, "net_accuracy": 0.22266401590457258}
mmlu_STEM/few_shot_1/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.23.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 429, "2": 336, "3": 0}, "net_accuracy": 0.25347912524850896}
mmlu_STEM/few_shot_1/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_blocks.24.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 347, "1": 26, "2": 341, "3": 24}, "net_accuracy": 0.24453280318091453}
mmlu_STEM/few_shot_1/EleutherAI_gpt-j-6B/last_token_relative_depth_100/accuracy_ln_final.hook_normalized.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 108, "1": 256, "2": 347, "3": 115}, "net_accuracy": 0.2736911862160371}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.0.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
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_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.0.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
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_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.1.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
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_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.1.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 176, "1": 0, "2": 0, "3": 647}, "net_accuracy": 0.2726971504307488}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.10.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 4, "1": 0, "2": 760, "3": 17}, "net_accuracy": 0.25878064943671303}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.10.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 613, "1": 0, "2": 0, "3": 39}, "net_accuracy": 0.21603711066931744}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.11.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 0, "2": 783, "3": 0}, "net_accuracy": 0.25944333996023855}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.11.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 185, "1": 159, "2": 171, "3": 248}, "net_accuracy": 0.25281643472498344}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.2.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 644, "1": 0, "2": 0, "3": 0}, "net_accuracy": 0.21338634857521538}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.2.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 1, "2": 773, "3": 0}, "net_accuracy": 0.25646123260437376}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.3.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 642, "1": 0, "2": 0, "3": 6}, "net_accuracy": 0.2147117296222664}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.3.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 1, "2": 776, "3": 0}, "net_accuracy": 0.25745526838966204}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.4.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 209, "1": 433, "2": 0, "3": 16}, "net_accuracy": 0.21802518223989398}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.5.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 7, "1": 8, "2": 579, "3": 214}, "net_accuracy": 0.26772697150430746}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.5.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 499, "1": 0, "2": 0, "3": 192}, "net_accuracy": 0.22895957587806495}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.6.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 34, "2": 274, "3": 514}, "net_accuracy": 0.27236580516898606}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.6.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
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_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.7.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 5, "2": 776, "3": 0}, "net_accuracy": 0.25878064943671303}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.7.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
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_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.8.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 14, "1": 704, "2": 0, "3": 0}, "net_accuracy": 0.2379058979456594}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.8.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 644, "1": 0, "2": 0, "3": 0}, "net_accuracy": 0.21338634857521538}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.9.hook_mlp_out.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 0, "1": 0, "2": 783, "3": 0}, "net_accuracy": 0.25944333996023855}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_blocks.9.hook_resid_post.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 641, "1": 2, "2": 0, "3": 2}, "net_accuracy": 0.21371769383697814}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_ln_final.hook_normalized.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 185, "1": 159, "2": 171, "3": 248}, "net_accuracy": 0.25281643472498344}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/accuracy_output.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"total_per_class": {"0": 645, "1": 719, "2": 783, "3": 871}, "correct_per_class": {"0": 185, "1": 159, "2": 171, "3": 248}, "net_accuracy": 0.25281643472498344}
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.0.hook_mlp_out.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.0.hook_resid_post.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.1.hook_mlp_out.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.1.hook_resid_post.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.10.hook_mlp_out.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.10.hook_resid_post.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.11.hook_mlp_out.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.11.hook_resid_post.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.2.hook_mlp_out.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.2.hook_resid_post.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.3.hook_mlp_out.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.3.hook_resid_post.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.4.hook_mlp_out.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.4.hook_resid_post.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.5.hook_mlp_out.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.5.hook_resid_post.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.6.hook_mlp_out.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.6.hook_resid_post.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.7.hook_mlp_out.csv ADDED
The diff for this file is too large to render. See raw diff
 
mmlu_STEM/few_shot_1/EleutherAI_gpt-neo-125M/last_token_relative_depth_100/intrinsic/mle_local_dims/blocks.7.hook_resid_post.csv ADDED
The diff for this file is too large to render. See raw diff