INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 Dudley

-0.77

imov

-0.73

izen

-0.70

lov

-0.69

verty

-0.67

iev

-0.66

 Kazakh

-0.66

éĩ

-0.65

ãĥīãĥ©ãĤ´ãĥ³

-0.64

itably

-0.64

POSITIVE LOGITS

 regress

0.79

esville

0.71

hift

0.69

iggurat

0.69

gments

0.68

liga

0.65

 ferment

0.65

hyde

0.63

 defect

0.63

 malfunction

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.