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

 Liberties

-0.78

aceous

-0.72

owsky

-0.71

Quote

-0.68

Cole

-0.68

uta

-0.66

ilogy

-0.65

Ly

-0.65

ACY

-0.64

root

-0.63

POSITIVE LOGITS

 autom

0.71

 Guang

0.70

 feared

0.69

 manufact

0.66

cedented

0.64

 chau

0.63

 demoral

0.62

 hypers

0.62

ax

0.61

sembly

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.