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

ogg

-0.74

packs

-0.73

UGE

-0.72

utz

-0.72

vier

-0.69

ugg

-0.69

pee

-0.69

fy

-0.68

iola

-0.68

dl

-0.66

POSITIVE LOGITS

gdala

0.70

ãĤ©

0.66

 commission

0.64

eteria

0.63

forth

0.63

moil

0.63

 remod

0.62

Nay

0.61

 Marriott

0.61

 Mellon

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.