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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

appropriate

-0.73

 Hurricanes

-0.71

eligible

-0.71

xious

-0.71

vae

-0.70

osphere

-0.68

ious

-0.68

etus

-0.68

externalActionCode

-0.68

bage

-0.66

POSITIVE LOGITS

WER

0.69

 stamp

0.68

rul

0.68

 estab

0.64

areth

0.63

 Stamp

0.60

 stamps

0.57

atem

0.57

EMS

0.57

pg

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.