INDEX

Explanations

titles and roles associated with former officials and political figures

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

former titles and professions

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

contentLoaded

-0.59

 NSCoder

-0.56

rawDesc

-0.53

 Comprometido

-0.50

tualmente

-0.47

ésult

-0.46

 pouvoit

-0.46

 saveiro

-0.45

erçe

-0.44

 câte

-0.44

POSITIVE LOGITS

 former

0.58

########.

0.56

 Former

0.53

Former

0.52

former

0.45

 formerly

0.42

endregion

0.42

:+:

0.39

 ehemaligen

0.39

 previous

0.37

Activations Density 0.035%

titles and roles associated with former officials and political figures

former titles and professions

No Comments

No Known Activations