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

oots

-0.80

ooting

-0.74

resses

-0.70

otton

-0.69

apon

-0.69

olor

-0.69

ressing

-0.66

olitan

-0.66

EStream

-0.65

apons

-0.65

POSITIVE LOGITS

ships

0.73

VALUE

0.73

ship

0.71

BRE

0.69

ray

0.68

Marginal

0.66

dies

0.65

english

0.63

STATE

0.63

Journal

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.