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

 halves

-0.72

 numer

-0.71

 mirrors

-0.67

 mirrored

-0.65

planes

-0.64

 Suns

-0.63

yrs

-0.63

 leagues

-0.63

 apes

-0.63

gir

-0.62

POSITIVE LOGITS

imum

0.85

ãĥ¼ãĥĨãĤ£

0.74

istor

0.71

Marginal

0.70

DEV

0.70

ailability

0.70

VD

0.69

thia

0.68

Brien

0.67

idable

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.