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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

POS

-0.70

gon

-0.67

wana

-0.65

gments

-0.65

tick

-0.64

 Layer

-0.61

 Morning

-0.60

irie

-0.59

ense

-0.58

itives

-0.58

POSITIVE LOGITS

audi

0.78

uthor

0.74

oped

0.69

 compr

0.68

hetti

0.66

weeney

0.65

igor

0.65

thumbnails

0.64

igers

0.63

bernatorial

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.