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

 autorytatywna

-0.44

OGND

-0.42

 "..\..\..\

-0.40

 Moscú

-0.39

__::

-0.39

 ſind

-0.39

retudo

-0.38

kheim

-0.38

 parís

-0.38

forChild

-0.36

POSITIVE LOGITS

 للاسماء

0.56

ViewFeatures

0.54

 oprot

0.54

LookAnd

0.54

 <<<<<<<<<<<<<<

0.50

 Chwiliwch

0.48

UnusedPrivate

0.47

 pinulongan

0.47

richTextPanel

0.47

WithIdentifier

0.45

Activations Density 0.000%

No Known Activations

This feature has no known activations.