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

RequiresApi

-0.49

Slf

-0.47

出版年

-0.47

CreateMap

-0.46

APN

-0.44

PageRoute

-0.44

█

-0.43

PhysRev

-0.43

/**

-0.42

 pizza

-0.42

POSITIVE LOGITS

ContextHolder

4.34

ScopeManager

0.72

 EconPapers

0.66

LookAnd

0.56

GOTREF

0.56

Context

0.52

ObjectContext

0.51

 GenerationType

0.50

AccessorTable

0.50

RuleContext

0.50

Activations Density 0.000%

No Known Activations

This feature has no known activations.