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

opers

-0.91

zb

-0.87

usk

-0.81

illard

-0.80

amon

-0.80

ust

-0.76

ustration

-0.76

zen

-0.75

ank

-0.73

ivo

-0.71

POSITIVE LOGITS

 worth

0.84

 liking

0.75

 soluble

0.67

 Magicka

0.67

 Frie

0.66

 wallets

0.66

Sly

0.65

 Souls

0.65

 costing

0.64

Wiz

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.