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

TypedDataSet

-0.67

+#+#

-0.66

AndroidJUnit

-0.56

HomeAsUpEnabled

-0.56

 informée

-0.54

 MainAxisSize

-0.52

 AttributeSet

-0.48

ésult

-0.48

 ModelExpression

-0.47

 BorderSide

-0.47

POSITIVE LOGITS

{§

0.37

Ecotoxicity

0.37

ímetros

0.36

səhifə

0.35

 tárgy

0.34

Tuy

0.33

sda

0.33

{\$

0.32

 nhiêu

0.32

msgstr

0.32

Activations Density 0.000%

No Known Activations

This feature has no known activations.