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

>{@

-0.89

WarningLevel

-0.69

GEBURTS

-0.40

Referanser

-0.35

KURZBESCHREIBUNG

-0.33

 anzi

-0.32

 infatti

-0.29

asikan

-0.28

jména

-0.28

 holand

-0.27

POSITIVE LOGITS

'{@

4.34

ember

1.24

 nahilalakip

0.80

 Alembic

0.78

 betweenstory

0.69

EndGlobalSection

0.68

 "..\..\..\

0.68

+#+

0.67

FunctionFlags

0.67

EMBER

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.