INDEX

Explanations

references to reading, hearing, or seeing information

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

found or heard somewhere

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

jao

-0.31

 cielos

-0.31

 sitä

-0.30

vertida

-0.28

 akhirnya

-0.28

 dificuldade

-0.28

 sonunda

-0.28

 življen

-0.28

avras

-0.28

 jälkeen

-0.27

POSITIVE LOGITS

ValueStyle

0.95

Personensuche

0.77

󠁿

0.75

IsMutable

0.73

extAlignment

0.68

 ویکی‌آمباردا

0.68

queryInterface

0.64

istoitu

0.62

:✨

0.62

EndInit

0.61

Activations Density 0.107%

references to reading, hearing, or seeing information

found or heard somewhere

No Comments

No Known Activations