INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.08

1:0.04

2:0.08

3:0.08

4:0.07

5:0.08

6:0.07

7:0.10

8:0.08

9:0.09

10:0.08

11:0.09

Negative Logits

gear

-1.67

craft

-1.65

among

-1.64

 attire

-1.58

shell

-1.57

 adherent

-1.51

keeper

-1.49

 accompanying

-1.48

Activ

-1.48

hel

-1.47

POSITIVE LOGITS

 Wonderland

1.79

 Conclusion

1.63

AAAAAAAA

1.60

Tan

1.58

 Winner

1.55

 Vale

1.53

────

1.51

ploma

1.49

rar

1.48

 Loan

1.46

Activations Density 0.000%

No Known Activations

This feature has no known activations.