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

rador

-0.69

ensor

-0.65

 Cabinet

-0.65

aling

-0.65

 Caps

-0.65

angular

-0.65

gie

-0.65

FSA

-0.64

gee

-0.63

haar

-0.63

POSITIVE LOGITS

orld

0.75

isode

0.75

 horr

0.67

 anew

0.66

OPLE

0.66

veyard

0.65

 dissatisf

0.64

song

0.64

eworld

0.63

onis

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.