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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

orah

-0.74

zos

-0.67

intosh

-0.64

mark

-0.63

 Advantage

-0.63

emis

-0.62

front

-0.62

true

-0.61

æµ

-0.59

Haw

-0.59

POSITIVE LOGITS

izoph

0.77

 cumbers

0.73

idates

0.73

ciating

0.72

accompan

0.70

 nomine

0.68

ADVERTISEMENT

0.66

 fateful

0.66

ovie

0.63

 carbohyd

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.