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

 largeDownload

-0.75

catentry

-0.72

orbit

-0.69

 ultraviolet

-0.67

otin

-0.67

 galleries

-0.64

involved

-0.64

ophical

-0.63

 âĶľ

-0.63

 rooft

-0.63

POSITIVE LOGITS

Force

0.73

xit

0.67

nings

0.63

ÄŁ

0.63

irs

0.62

 jihad

0.62

 Emin

0.62

 Resurrection

0.62

hz

0.62

otte

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.