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

dry

-0.35

 Reif

-0.34

மை

-0.32

hit

-0.30

 tall

-0.30

=$((

-0.29

mid

-0.29

 ugly

-0.28

 legal

-0.28

 atop

-0.28

POSITIVE LOGITS

 NavController

5.53

NavController

1.70

 OnInit

1.02

 HomeController

0.87

StackNavigator

0.83

IntoConstraints

0.82

verwijspagina

0.81

 useNavigate

0.79

MemoryWarning

0.78

сылкі

0.77

Activations Density 0.000%

No Known Activations

This feature has no known activations.