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

urst

-0.91

ipeg

-0.78

iott

-0.78

astery

-0.77

terday

-0.76

itaire

-0.73

urrencies

-0.69

ĸļ

-0.69

 incent

-0.66

olicy

-0.65

POSITIVE LOGITS

0.83

å°Ĩ

0.72

birds

0.66

Bruce

0.66

PLE

0.66

Bus

0.65

GV

0.64

ãĢı

0.64

 Savior

0.64

SC

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.