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

stare

-0.48

leps

-0.45

ault

-0.44

 stare

-0.44

roll

-0.41

ತ

-0.40

elope

-0.40

prototype

-0.40

 digress

-0.40

PARAMETER

-0.39

POSITIVE LOGITS

wsgi

4.00

asgi

1.95

WSGI

1.75

dafx

0.56

 Roskov

0.53

 Administrativna

0.52

jspb

0.52

SpringBootTest

0.52

 ComVisible

0.52

 jspb

0.51

Activations Density 0.000%

No Known Activations

This feature has no known activations.