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

IZONT

-0.40

 appraisals

-0.40

 yourself

-0.39

unyai

-0.39

meteor

-0.39

raisals

-0.39

 бывает

-0.38

 aprendiz

-0.38

TargetException

-0.38

arte

-0.38

POSITIVE LOGITS

prnewswire

4.81

twimg

1.46

PRNewswire

1.35

newswire

0.99

PRWEB

0.92

 Newswire

0.83

Newswire

0.78

SequentialGroup

0.74

DotNetBar

0.64

bootstrapcdn

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.