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

EStream

-0.73

udeb

-0.72

 Puppet

-0.71

 Narr

-0.70

avement

-0.70

andowski

-0.70

ulner

-0.67

hoe

-0.66

 Fedora

-0.65

æĪ¦

-0.65

POSITIVE LOGITS

com

1.34

org

0.70

joining

0.69

 sorcery

0.67

COM

0.67

cigarettes

0.66

 millennium

0.66

 cigarettes

0.64

com

0.64

tv

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.