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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Vehicle

-0.69

AW

-0.65

ESA

-0.65

urat

-0.64

Surv

-0.64

WE

-0.63

Detailed

-0.63

GST

-0.63

 pelvic

-0.62

zh

-0.62

POSITIVE LOGITS

ãĤ´ãĥ³

0.74

¬¼

0.73

agos

0.69

ģĸ

0.69

[|

0.67

Ĺ

0.67

lished

0.67

Ĥª

0.66

Ń·

0.66

«ĺ

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.