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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ã§a

-0.70

inyl

-0.67

 âĵĺ

-0.59

Xan

-0.59

Grey

-0.59

 gamma

-0.59

).

-0.57

 DRAG

-0.57

VG

-0.57

 Albion

-0.57

POSITIVE LOGITS

 today

0.76

prints

0.70

 Hirosh

0.64

hend

0.60

itect

0.60

uthor

0.60

hid

0.59

writ

0.59

 elig

0.59

à¥

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.