INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

folk

-0.71

 evenly

-0.67

 Pixie

-0.66

 Cole

-0.66

 Prince

-0.65

 Rebell

-0.64

soType

-0.63

 Flor

-0.62

Fey

-0.61

 scouting

-0.60

POSITIVE LOGITS

phabet

0.98

mx

0.80

Mach

0.78

trak

0.76

ZI

0.74

Reviewer

0.73

Â·Â·

0.73

eday

0.72

NK

0.71

orthern

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.