INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

soDeliveryDate

-0.88

isSpecialOrderable

-0.83

perture

-0.80

 attRot

-0.74

é¾įå¥ĳå£«

-0.68

éĹĺ

-0.68

LU

-0.67

ãĥĵ

-0.67

è¦ļéĨĴ

-0.66

TEXTURE

-0.65

POSITIVE LOGITS

mone

0.69

ono

0.69

 agre

0.65

agara

0.64

oun

0.63

deen

0.61

acer

0.60

 unres

0.59

ominated

0.58

 Worse

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.