INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

itatively

-0.86

 awoken

-0.85

occup

-0.81

REDACTED

-0.76

acca

-0.74

Ã±a

-0.73

iens

-0.73

dayName

-0.72

illation

-0.71

ciating

-0.70

POSITIVE LOGITS

 Founders

0.76

 captcha

0.75

 Compass

0.71

 Fram

0.69

Duo

0.65

 Friends

0.65

DeL

0.64

Sik

0.64

 Fork

0.63

Tud

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.