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

EndInit

-0.53

 препратки

-0.51

CGRectMake

-0.47

 QLabel

-0.44

eseorang

-0.44

 prosz

-0.44

 arşivlendi

-0.44

 Learner

-0.44

comings

-0.43

 Arenas

-0.43

POSITIVE LOGITS

tagHelperRunner

0.54

ⓧ

0.45

usercontent

0.43

blurRadius

0.39

NUC

0.37

 Archiproducts

0.36

Naw

0.35

所示

0.34

 Numerade

0.34

adaptiveStyles

0.33

Activations Density 0.000%

No Known Activations

This feature has no known activations.