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

owl

-0.75

rant

-0.70

 wart

-0.66

raising

-0.65

ram

-0.64

PLE

-0.63

ording

-0.62

 scrolling

-0.61

boa

-0.61

CLA

-0.60

POSITIVE LOGITS

 horizont

0.76

 indo

0.68

ãĤ¼ãĤ¦ãĤ¹

0.67

 Locke

0.66

 Niet

0.66

icts

0.66

 Nare

0.65

istry

0.65

 Swift

0.64

 Roose

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.