INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

Ã¼r

-0.80

channelAvailability

-0.74

aeper

-0.73

ï¿½ï¿½

-0.71

arrass

-0.70

ģ

-0.69

umper

-0.68

uch

-0.68

aceutical

-0.66

mega

-0.66

POSITIVE LOGITS

sov

0.80

bane

0.79

breeding

0.75

entin

0.74

worms

0.70

 inspir

0.67

 warranties

0.66

aneous

0.65

nc

0.65

lando

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.