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

 paddle

-0.67

 Bhar

-0.62

Wr

-0.62

 whistle

-0.61

ethy

-0.60

 magically

-0.60

ple

-0.57

due

-0.57

 haun

-0.57

 borrower

-0.56

POSITIVE LOGITS

Advertisements

0.88

CLASSIFIED

0.76

 apartheid

0.70

lled

0.70

advertisement

0.69

mosp

0.68

bris

0.68

acements

0.65

imov

0.65

Draft

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.