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

 slutt

-0.50

 exportación

-0.45

 république

-0.43

 óptica

-0.43

 turística

-0.42

 paraître

-0.42

empêcher

-0.41

muñeca

-0.41

 mensuales

-0.41

 anuales

-0.41

POSITIVE LOGITS

inflater

5.22

 inflater

1.40

Inflater

1.11

FormTagHelper

0.84

LayoutInflater

0.83

 infla

0.80

Infl

0.76

 Inflate

0.76

inflate

0.75

MenuInflater

0.73

Activations Density 0.000%

No Known Activations

This feature has no known activations.