INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

etus

-0.75

 Sutherland

-0.74

tsky

-0.68

VG

-0.66

CoC

-0.66

 Loading

-0.65

gur

-0.65

Home

-0.64

 Parenthood

-0.63

POSITIVE LOGITS

tnc

0.74

stood

0.68

jog

0.67

 colle

0.64

matic

0.64

 Scare

0.62

oooooooo

0.62

 rook

0.61

 initiative

0.61

advert

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.