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

 purse

-0.67

uries

-0.66

 pockets

-0.65

pocket

-0.65

iku

-0.64

 Metropolitan

-0.63

isi

-0.62

 Flavoring

-0.62

 Marse

-0.60

ã

-0.60

POSITIVE LOGITS

 pestic

0.88

 horm

0.80

 reconc

0.78

 Scientist

0.68

 timed

0.67

 artif

0.66

 mathemat

0.65

 Stead

0.65

 overfl

0.65

ĵĺ

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.