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

oÄŁ

-0.99

sembly

-0.87

 largeDownload

-0.74

veyard

-0.73

 Magicka

-0.73

¥µ

-0.71

Ã±a

-0.70

 larvae

-0.68

zona

-0.68

 Pengu

-0.66

POSITIVE LOGITS

Quote

0.81

ITAL

0.81

Shares

0.80

Hill

0.79

OY

0.76

Norm

0.75

Warren

0.75

Subject

0.74

Trust

0.73

Relations

0.71

Activations Density 0.000%

No Known Activations

This feature has no known activations.