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

 istg

-0.81

"]=>

-0.68

OM

-0.67

 discrep

-0.65

 PHOTO

-0.64

 administ

-0.64

 Programme

-0.63

 Authorization

-0.61

 Atmosp

-0.60

 ener

-0.59

POSITIVE LOGITS

river

0.80

bernatorial

0.79

uncle

0.76

orum

0.71

insula

0.67

geist

0.67

lund

0.67

Mah

0.66

afort

0.66

behavior

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.