INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ãĥĢ

-0.78

 Annie

-0.68

ãĤ³

-0.67

Sax

-0.67

ãĥŀ

-0.67

 Amer

-0.66

 Atlas

-0.66

ãĥ¼ãĥ«

-0.65

ãĤ´ãĥ³

-0.64

":"/

-0.63

POSITIVE LOGITS

ingers

0.77

hops

0.74

selves

0.73

quet

0.68

cru

0.68

hou

0.65

cru

0.64

ongs

0.63

senal

0.63

ittered

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.