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

 Rebell

-0.70

ances

-0.64

asers

-0.63

inson

-0.63

prints

-0.62

RIP

-0.62

 REPL

-0.62

 presidents

-0.61

 reps

-0.60

Aux

-0.59

POSITIVE LOGITS

hid

0.85

hig

0.74

ouf

0.74

advertising

0.73

vernment

0.73

ãĥĻ

0.69

vati

0.67

oS

0.67

%);

0.65

utsu

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.