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

Anonymous

-0.71

-0.69

 ILCS

-0.68

SOURCE

-0.68

talk

-0.66

jab

-0.64

 diction

-0.60

Filename

-0.58

 antioxid

-0.57

 watts

-0.57

POSITIVE LOGITS

lass

0.78

sth

0.74

beh

0.70

ead

0.69

isch

0.68

UGE

0.67

atform

0.65

worldly

0.64

thouse

0.62

 Statue

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.