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

anguage

-0.81

ference

-0.77

sight

-0.75

hens

-0.70

orks

-0.70

hemy

-0.67

 clause

-0.66

oldown

-0.64

 brunch

-0.64

reads

-0.64

POSITIVE LOGITS

 relative

1.00

ãĤ©

0.83

 looking

0.75

Yad

0.68

ierrez

0.63

 REUTERS

0.62

Benz

0.61

 doomed

0.60

 indebted

0.60

itably

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.