INDEX

Explanations

No Explanations Found

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 estekak

-0.46

 oprot

-0.43

CONCLUSIONES

-0.42

 uVar

-0.42

ṁ

-0.41

abestanden

-0.41

 iconTwitter

-0.41

InputTagHelper

-0.40

})()

-0.39

Ҭ

-0.39

POSITIVE LOGITS

verlet

0.39

 calculées

0.39

WebVitals

0.37

 Feind

0.37

tagHelperRunner

0.36

 ANGELES

0.35

 jezelf

0.35

 planific

0.35

烃

0.34

 sterili

0.34

Activations Density 0.000%

No Known Activations

This feature has no known activations.