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

0:0.09

1:0.08

2:0.07

3:0.08

4:0.07

5:0.09

6:0.07

7:0.08

8:0.08

9:0.09

10:0.08

11:0.07

Negative Logits

 gaps

-1.68

 discrep

-1.65

gap

-1.62

��

-1.61

 resulting

-1.51

tions

-1.51

 probability

-1.50

 differential

-1.48

AME

-1.47

 disparity

-1.47

POSITIVE LOGITS

terday

1.76

 Thurs

1.66

stru

1.65

 Sundays

1.64

 Fridays

1.60

 Saturdays

1.56

 twent

1.54

 Travels

1.49

 Tues

1.47

past

1.47

Activations Density 0.000%

No Known Activations

This feature has no known activations.