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

sym

-0.67

udi

-0.63

iae

-0.63

iate

-0.62

uve

-0.61

VIS

-0.60

 Kali

-0.60

 Bengal

-0.59

iations

-0.59

aneously

-0.59

POSITIVE LOGITS

sheet

0.86

 ç¥ŀ

0.83

ãĥ¯ãĥ³

0.79

essee

0.73

 compan

0.72

oday

0.71

hots

0.70

etting

0.69

Unknown

0.67

riks

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.