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

rontal

-0.74

 neighb

-0.73

 Seym

-0.70

ucci

-0.68

obos

-0.67

seen

-0.66

ß

-0.66

MpServer

-0.66

anian

-0.65

ansson

-0.65

POSITIVE LOGITS

 Trop

0.71

 Able

0.70

 interstate

0.68

 Sturgeon

0.67

incial

0.66

assetsadobe

0.64

 Liter

0.62

 Sierra

0.62

cott

0.59

 Sask

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.