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

 Wikiseite

-0.47

expandindo

-0.42

 pronti

-0.41

 indépendante

-0.40

 psicológica

-0.40

 Stirn

-0.39

 لينك

-0.39

 muerta

-0.39

 independientes

-0.37

 geométrico

-0.37

POSITIVE LOGITS

 TType

1.34

 CURIAM

0.52

szy

0.51

Composable

0.49

 iprot

0.48

titleMargin

0.46

 Superhost

0.46

bootstrapcdn

0.45

modelBuilder

0.44

setts

0.44

Activations Density 0.000%

No Known Activations

This feature has no known activations.