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

 Administrativna

-1.03

$_['

-0.68

 surla

-0.57

EndGlobalSection

-0.56

 <>",

-0.56

 Biôgrafia

-0.54

TokenNameLPAREN

-0.54

Scénario

-0.54

 createSprite

-0.53

WriteTagHelper

-0.52

POSITIVE LOGITS

 katkıda

0.64

 innovation

0.37

FieldNumber

0.33

Referências

0.31

 TestCase

0.31

促

0.31

WritableDatabase

0.30

 excellence

0.30

 Vikipedi

0.30

 innovación

0.30

Activations Density 0.000%

No Known Activations

This feature has no known activations.