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

ħ

-0.62

HB

-0.61

Ry

-0.61

quarter

-0.57

————

-0.57

ND

-0.57

ITION

-0.56

 Sabha

-0.56

SourceFile

-0.55

 supplemented

-0.55

POSITIVE LOGITS

arcity

0.76

ograph

0.75

upid

0.72

udic

0.72

netflix

0.71

itudes

0.69

phrase

0.68

ificant

0.68

icable

0.67

antes

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.