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

gabe

-0.48

 zoos

-0.46

Giving

-0.45

gegeben

-0.44

yv

-0.43

phosa

-0.42

DataAnnotations

-0.41

 Muhamma

-0.41

 BeautifulSoup

-0.41

resented

-0.41

POSITIVE LOGITS

BeforeClass

2.78

AfterClass

2.14

BeforeAll

1.20

BeforeMethod

1.05

PostConstruct

0.77

beforeAll

0.71

GEBURTSDATUM

0.59

classmethod

0.56

 betweenstory

0.54

 BEFORE

0.51

Activations Density 0.000%

No Known Activations

This feature has no known activations.