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

UnusedPrivate

-0.59

AutoField

-0.54

 utafitiHapana

-0.53

تقاوى

-0.50

 propOrder

-0.48

Дереккөздер

-0.47

Попис

-0.46

Искәрмәләр

-0.45

 AssemblyTitle

-0.45

ğraf

-0.44

POSITIVE LOGITS

 desertcart

1.51

desertcart

0.58

StructEnd

0.45

cartão

0.44

AutoModerator

0.42

 sostenibilidad

0.41

 marketplaces

0.40

esticides

0.40

 gubern

0.39

Kart

0.39

Activations Density 0.000%

No Known Activations

This feature has no known activations.