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

trak

-1.07

HUD

-0.97

 ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³

-0.92

hare

-0.85

wcs

-0.78

govtrack

-0.76

yip

-0.71

phis

-0.70

hig

-0.70

 UCHIJ

-0.70

POSITIVE LOGITS

olate

0.66

 acquaintance

0.66

 Lies

0.65

 Chronicles

0.63

Dys

0.62

 enorm

0.62

chen

0.60

 Monteneg

0.60

 herself

0.60

 angle

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.