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

vertisements

-0.71

 Democr

-0.69

 arisen

-0.66

 malf

-0.66

 telev

-0.66

 exoner

-0.66

 frivol

-0.66

ities

-0.65

izoph

-0.65

éĸ

-0.65

POSITIVE LOGITS

GGGG

0.79

love

0.75

Gender

0.72

orius

0.72

brance

0.72

NRS

0.71

Artist

0.70

atti

0.70

NUM

0.69

Joshua

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.