INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Bioaccumulative

-0.46

queryInterface

-0.43

GHIJKLM

-0.41

 consectetur

-0.40

ագրություններ

-0.40

 Administrativna

-0.39

Бахар

-0.39

 isol

-0.38

Geplaatst

-0.38

 Roskov

-0.37

POSITIVE LOGITS

 betweenstory

0.67

 plufieurs

0.60

 dieß

0.57

 Monfieur

0.50

 dezelve

0.49

 ſont

0.48

 jTextField

0.48

ſelf

0.48

 Houſe

0.47

GEBURTSDATUM

0.47

Activations Density 0.000%

No Known Activations

This feature has no known activations.