INDEX

Explanations

No Explanations Found

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

RenderAtEndOf

-0.41

tagext

-0.40

 prop

-0.35

 hypothetical

-0.34

 stát

-0.33

ReusableCell

-0.32

two

-0.32

 Hukum

-0.32

FormUrlEncoded

-0.32

 verità

-0.31

POSITIVE LOGITS

ArrowToggle

1.70

NameInMap

0.54

 يتيمه

0.53

oredCriteria

0.53

rungsseite

0.51

 continúas

0.51

Chham

0.49

########.

0.49

AndEndTag

0.48

 informée

0.48

Activations Density 0.000%

No Known Activations

This feature has no known activations.