INDEX

Explanations

titles and roles related to corporate hierarchy and management

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

officials making statements

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

 transfieras

-0.69

 abstrato

-0.56

transQ

-0.56

/*

-0.56

<?

-0.52

 للمعارف

-0.52

 zijne

-0.51

SequentialGroup

-0.51

 mijne

-0.51

 Италијани

-0.50

POSITIVE LOGITS

 spokesman

0.45

adi

0.44

AddHtmlAttribute

0.43

said

0.42

 said

0.41

她说

0.41

Selfer

0.40

 secretary

0.40

 Hass

0.40

 explained

0.39

Activations Density 0.013%

titles and roles related to corporate hierarchy and management

officials making statements

No Comments

No Known Activations