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

Component

-0.66

Ĥª

-0.65

Progress

-0.63

 Territories

-0.63

 schematic

-0.62

bands

-0.62

oooooooo

-0.62

ãĥ¼ãĥĨãĤ£

-0.59

displayText

-0.57

 marched

-0.57

POSITIVE LOGITS

rent

0.73

ibur

0.73

agu

0.71

hedon

0.70

wine

0.69

isexual

0.68

umers

0.68

 Noir

0.66

anamo

0.65

ric

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.