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

has

-0.33

el

-0.33

hex

-0.32

get

-0.32

 invited

-0.31

czy

-0.30

my

-0.30

 online

-0.30

🟠

-0.29

set

-0.29

POSITIVE LOGITS

apimachinery

6.09

 metav

1.01

mybatisplus

0.96

 defaultstate

0.94

 surla

0.90

parsedMessage

0.89

ChildScrollView

0.88

 kasarigan

0.87

CppMethod

0.87

 ProtoMessage

0.87

Activations Density 0.000%

No Known Activations

This feature has no known activations.