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

kaya

-0.82

patrick

-0.80

aq

-0.68

 glac

-0.65

 Choi

-0.64

 Kuwait

-0.64

å¤©

-0.64

Kuh

-0.63

 Buchanan

-0.62

bah

-0.62

POSITIVE LOGITS

isSpecialOrderable

0.69

ware

0.68

 Chrome

0.68

Frames

0.64

umbered

0.62

 Classification

0.62

 stitches

0.62

parable

0.61

Mask

0.60

 browsers

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.