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

GOTREF

-0.41

openhauer

-0.38

 Maddox

-0.34

tuigen

-0.34

resized

-0.34

}>

-0.33

 geç

-0.33

 merely

-0.33

yaxis

-0.32

')}}">

-0.32

POSITIVE LOGITS



0.96

Obituary

0.81

IndentedString

0.74

Carriera

0.69

complexContent

0.66

 autorytatywna

0.65

Tembelea

0.62

UARY

0.60

出版年

0.59

VersionUID

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.