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

OH

-0.71

 ages

-0.68

[];

-0.66

cov

-0.66

ozyg

-0.66

haar

-0.65

 deed

-0.64

BILITIES

-0.64

arks

-0.64

GH

-0.63

POSITIVE LOGITS

 Metatron

0.75

 Faith

0.71

athe

0.69

Publisher

0.66

tz

0.64

adan

0.64

gow

0.64

 EntityItem

0.64

Wiki

0.63

Faith

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.