INDEX

Explanations

words and phrases related to engineering and design processes

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

and words

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

 male

-0.33

 vrou

-0.29

 condizione

-0.29

 poveznice

-0.29

fromnode

-0.28

 حا

-0.28

äch

-0.28

goc

-0.28

 female

-0.28

 га

-0.27

POSITIVE LOGITS

 ویکی‌آمباردا

0.65

NameInMap

0.63

الحياه

0.63

 estekak

0.60

 Normdatei

0.57

ContentAlignment

0.56

 onPostExecute

0.55

dafx

0.54

 Italijani

0.54

BrowserModule

0.53

Activations Density 0.022%

words and phrases related to engineering and design processes

and words

No Comments

No Known Activations