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

chain

-0.42

rome

-0.41

gene

-0.40

park

-0.38

ous

-0.38

utt

-0.37

 kost

-0.37

babel

-0.37

sea

-0.37

vfs

-0.36

POSITIVE LOGITS

afficheront

0.74

 EconPapers

0.73

setVerticalGroup

0.72

adaptiveStyles

0.71

parsedMessage

0.68

Билгалдахарш

0.66

ImageContext

0.61

webElementXpaths

0.60

 gyhoeddwyd

0.59

WebElementEntity

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.