INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

urdue

-0.79

swick

-0.78

cham

-0.75

ORTS

-0.74

conn

-0.71

Draft

-0.68

_-

-0.68

|--

-0.67

lehem

-0.65

PLIED

-0.64

POSITIVE LOGITS

unal

0.77

tube

0.72

 pals

0.71

 bombard

0.66

 activ

0.65

 youngster

0.63

ÃĥÃĤ

0.63

 youngsters

0.62

 maniac

0.62

 arsen

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.