INDEX

Explanations

articles and specific letters or numbers in text

oai_token-act-pair · gpt-4o-mini Triggered by @bot

articles and specific entities

np_acts-logits-general · gemini-2.5-flash-lite

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

dición

-0.39

apeake

-0.39

IntoConstraints

-0.38

didSet

-0.38

falgar

-0.37

PageContext

-0.37

centes

-0.36

دانشنامهٔ

-0.36

 corrientes

-0.36

 amarillas

-0.36

POSITIVE LOGITS

ंदीखरीदारी

0.56

 Италијани

0.52

Personensuche

0.50

fjspx

0.49

 مرئيه

0.48

 kohdetta

0.46

 Taktlose

0.44

 OMITBAD

0.43

期刊论文

0.41

principalTable

0.40

Activations Density 0.022%

articles and specific letters or numbers in text

articles and specific entities

No Comments

No Known Activations