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

 taught

-0.41

DataPropertyName

-0.40

 hero

-0.39

yar

-0.39

dangerous

-0.36

enf

-0.36

FR

-0.35

 šķ

-0.35

pos

-0.35

back

-0.35

POSITIVE LOGITS

oredCriteria

1.41

parsedMessage

0.76

 يتيمه

0.67

ंदीखरीदारी

0.65

Билгалдахарш

0.63

 Roskov

0.63

rungsseite

0.62

әрмәләр

0.60

verwijspagina

0.60

хьтан

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.