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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ネルギー

-0.38

 eléctricos

-0.36

 elétrico

-0.35

 JsonResult

-0.35

 bénéfice

-0.35

ThroughAttribute

-0.33

 aéreo

-0.33

 Werbung

-0.33

ngrx

-0.33

 eléctrico

-0.32

POSITIVE LOGITS

VersionUID

6.34

OGND

0.90

WriteTagHelper

0.88

 AssemblyCulture

0.86

 nahilalakip

0.84

httphttps

0.79

MigrationBuilder

0.78

 AssemblyTitle

0.76

AddTagHelper

0.75

 AssemblyVersion

0.74

Activations Density 0.000%

No Known Activations

This feature has no known activations.