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

Weak

-0.87

 Flavoring

-0.84

////////////////

-0.75

dds

-0.74

Tele

-0.73

unc

-0.72

tsy

-0.72

YR

-0.71

 guiName

-0.71

Î

-0.70

POSITIVE LOGITS

 Sakuya

0.70

 PARK

0.68

 Kelly

0.66

apsed

0.64

 cascade

0.64

 citations

0.63

 Barkley

0.62

 enriched

0.61

 Cardinals

0.60

 scrub

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.