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

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.76

hower

-0.76

æ©

-0.73

ãģ®å®

-0.72

icative

-0.66

ãĥ¼ãĥĨãĤ£

-0.66

thinkable

-0.65

hov

-0.65

DEL

-0.64

ãĤ¼ãĤ¦ãĤ¹

-0.64

POSITIVE LOGITS

rame

0.73

ouch

0.66

Â£

0.65

ross

0.64

rowing

0.64

products

0.63

mist

0.63

ritis

0.63

grave

0.62

rily

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.