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

Eur

-0.77

YP

-0.75

 Azerbai

-0.68

 Syri

-0.63

KR

-0.63

 authors

-0.63

 Juda

-0.62

 Shogun

-0.61

 diplom

-0.60

Zar

-0.60

POSITIVE LOGITS

ties

0.78

gal

0.78

achusetts

0.71

local

0.67

venth

0.65

utenant

0.64

served

0.64

eteenth

0.63

coon

0.62

eteen

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.