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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

PIN

-0.69

cule

-0.66

Ô

-0.66

âĦ¢:

-0.65

APP

-0.64

assium

-0.63

 luggage

-0.62

lim

-0.62

ileaks

-0.62

ãĥ¤

-0.61

POSITIVE LOGITS

xual

0.74

 Glover

0.63

Mew

0.63

 Dragons

0.61

hematic

0.60

uthor

0.60

 Greens

0.60

 Machina

0.59

DERR

0.58

 Cunning

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.