INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Quantity

-1.01

Mi

-0.81

Impl

-0.75

Avg

-0.73

catentry

-0.71

VERSION

-0.67

Pwr

-0.67

uphem

-0.66

PN

-0.66

NG

-0.66

POSITIVE LOGITS

ariat

0.86

bom

0.75

assi

0.68

inates

0.64

inic

0.63

alling

0.62

luent

0.61

emouth

0.60

 ruined

0.60

 displaced

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.