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

igslist

-0.90

ciating

-0.76

VIDIA

-0.69

 behavi

-0.67

 arose

-0.67

halla

-0.66

assic

-0.63

anooga

-0.62

voc

-0.61

wiki

-0.60

POSITIVE LOGITS

enegger

0.80

 upgr

0.73

saf

0.72

atility

0.69

atile

0.68

lings

0.68

rated

0.66

neath

0.65

obos

0.64

locked

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.