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

出版年

-0.65

 Infórmanos

-0.63

MLLoader

-0.61

complexContent

-0.59

ogical

-0.59

 useAppContext

-0.58

AddTagHelper

-0.57

 Exacts

-0.57

textAppearance

-0.56

 betweenstory

-0.54

POSITIVE LOGITS

TabIndex

0.52

 sonunda

0.38

 saites

0.37

WithMany

0.35

setViewportView

0.35

 nedeniyle

0.34

 Accesorios

0.33

不知

0.33

Kjelder

0.33

 Hierro

0.33

Activations Density 0.000%

No Known Activations

This feature has no known activations.