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

ReusableCell

-0.68

BindingResult

-0.56

StartTag

-0.53

 يتيمه

-0.53

ViewInit

-0.51

matchCondition

-0.50

tvguidetime

-0.45

instancetype

-0.44

stretchr

-0.44

javax

-0.44

POSITIVE LOGITS

hyperref

1.67

 hyperlink

0.57

Хьажоргаш

0.53

GEBURTSDATUM

0.52

 estekak

0.52

HomeAsUpEnabled

0.51

 gynhyrchwyd

0.50

 Италијани

0.47

 henvis

0.45

poptotic

0.44

Activations Density 0.000%

No Known Activations

This feature has no known activations.