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

rohr

-0.36

out

-0.33

mongoose

-0.32

 prosecution

-0.32

 ум

-0.31

 ferocious

-0.31

goto

-0.31

high

-0.31

en

-0.31

 stellar

-0.30

POSITIVE LOGITS

initState

0.96

 StatefulWidget

0.88

 createState

0.79

 StatelessWidget

0.73

 setState

0.71

 CURIAM

0.70

CppMethod

0.67

 ControllerBase

0.66

GOTREF

0.65

onViewCreated

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.