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

sted

-0.78

nesday

-0.76

amped

-0.75

elman

-0.73

 quarry

-0.68

 quar

-0.66

ignment

-0.65

 tenant

-0.65

near

-0.64

icz

-0.64

POSITIVE LOGITS

ãĤ©

0.91

 Cheong

0.87

Browser

0.78

0.76

ãĥĩãĤ£

0.69

amins

0.67

FTWARE

0.67

gamer

0.67

goers

0.67

 Delicious

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.