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

ickr

-0.73

irmation

-0.64

minster

-0.63

OUN

-0.61

 Russ

-0.61

 fulfillment

-0.61

PTS

-0.60

ontent

-0.60

 guiActiveUnfocused

-0.60

=#

-0.59

POSITIVE LOGITS

acial

0.67

eva

0.65

avorite

0.65

Ĥª

0.64

pread

0.64

 minions

0.63

ierrez

0.63

duction

0.62

ulton

0.62

 Cardinal

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.