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

tnc

-0.89

iew

-0.79

thren

-0.73

nown

-0.72

wings

-0.71

ublic

-0.70

soDeliveryDate

-0.69

bye

-0.68

isSpecialOrderable

-0.68

paren

-0.67

POSITIVE LOGITS

ventus

0.87

 Ogre

0.68

ãĥĸ

0.65

ãĥĦ

0.65

 Hodg

0.64

Ru

0.61

Hip

0.61

 Squid

0.60

 hits

0.60

Rid

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.