INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

=-=-=-=-=-=-=-=-

-0.75

staking

-0.72

foot

-0.66

 Barg

-0.64

DragonMagazine

-0.64

raph

-0.63

 foregoing

-0.60

 Ãľ

-0.59

âĹ¼

-0.56

NetMessage

-0.56

POSITIVE LOGITS

antha

0.78

iture

0.77

odore

0.70

ature

0.70

icious

0.69

IPS

0.68

EY

0.67

ashtra

0.66

itaire

0.65

Ĥª

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.