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

 Nept

-0.77

culosis

-0.77

é¾

-0.76

nces

-0.71

tions

-0.71

ukong

-0.71

erent

-0.70

ronics

-0.69

Sup

-0.68

DragonMagazine

-0.67

POSITIVE LOGITS

 therefore

0.81

 hence

0.79

 thence

0.74

 consequently

0.72

 thus

0.71

 mortar

0.68

0.63

 pounding

0.61

 bruised

0.60

 thirsty

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.