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

ﬅ

-0.47

UIAlert

-0.32

 το

-0.32

 yılı

-0.32

ậy

-0.32

 kiệm

-0.31

 homes

-0.30

el

-0.30

ನ್ನು

-0.30

 aquel

-0.30

POSITIVE LOGITS

BeginInit

6.09

EndInit

2.70

PerformLayout

1.25

SuspendLayout

1.21

GOTREF

1.02

InitVars

0.92

 initComponents

0.90

 nahilalakip

0.86

setVerticalGroup

0.84

 مشين

0.84

Activations Density 0.000%

No Known Activations

This feature has no known activations.