INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ãĥĸ

-0.76

glers

-0.73

 diam

-0.71

 Destroyer

-0.70

swick

-0.69

teness

-0.69

gob

-0.67

 Crim

-0.67

 Cumber

-0.66

bral

-0.65

POSITIVE LOGITS

 even

1.16

even

0.97

 EVEN

0.87

abetic

0.73

forward

0.69

 uncover

0.67

 tasting

0.66

yip

0.65

ellery

0.62

estic

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.