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

ĸļ

-0.68

 Mosque

-0.68

 Cheng

-0.65

 freezer

-0.64

 Zone

-0.62

ezvous

-0.61

abulary

-0.61

ulu

-0.59

 Malaysian

-0.59

insky

-0.58

POSITIVE LOGITS

 advoc

0.86

ired

0.75

change

0.70

ield

0.69

 holdings

0.67

leaf

0.66

utenberg

0.65

 toget

0.65

lete

0.65

 tradem

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.