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

MpServer

-0.79

ortality

-0.64

earable

-0.64

ADE

-0.64

Row

-0.63

 wavelength

-0.62

ierrez

-0.61

 notch

-0.61

66666666

-0.61

 designation

-0.60

POSITIVE LOGITS

edom

0.79

ateur

0.72

comed

0.72

itty

0.67

folk

0.67

isters

0.65

athe

0.65

podcast

0.64

inter

0.63

iqu

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.