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

ite

-0.38

↵↵

-0.36

 take

-0.35

 special

-0.35

Yes

-0.34

ing

-0.33

#+#

-0.33

-0.32

 free

-0.32

POSITIVE LOGITS

 estekak

7.53

OGND

3.20

 مشين

1.12

 propOrder

1.11

tagHelperRunner

1.09

 ligiloj

1.08

IntoConstraints

1.02

 esternos

1.00

expandindo

0.98

 يتيمه

0.97

Activations Density 0.000%

No Known Activations

This feature has no known activations.