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

und

-0.68

nos

-0.63

ryce

-0.60

âģ

-0.60

video

-0.59

bir

-0.59

opard

-0.58

mi

-0.57

gging

-0.56

cd

-0.56

POSITIVE LOGITS

acons

0.76

ients

0.75

ience

0.67

 Architects

0.66

 Aires

0.65

umbledore

0.64

 Canberra

0.64

 minions

0.63

enfranch

0.63

AUT

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.