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

 Kompon

-0.36

 vectorielle

-0.35

ถม

-0.34

saraba

-0.34

 islands

-0.33

 Vereinigte

-0.33

ائه

-0.33

 évent

-0.32

Cursor

-0.32

 peines

-0.32

POSITIVE LOGITS

detectChanges

6.00

SaveChangesAsync

0.92

getDoctrine

0.90

WriteBarrier

0.83

StoryboardSegue

0.82

 TestBed

0.82

IndentedString

0.82

printStackTrace

0.79

makeConstraints

0.78

TestingModule

0.76

Activations Density 0.000%

No Known Activations

This feature has no known activations.