INDEX

Explanations

titles and honorifics used to refer to individuals

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

titles preceding names

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

 TextInputType

-0.66

BeginContext

-0.66

ⓧ

-0.66

 BorderSide

-0.64

ItemBackground

-0.64

 snippetHide

-0.63

 ब्रेकडाउन

-0.62

 تضيفلها

-0.61

iNdEx

-0.61

BoxShadow

-0.60

POSITIVE LOGITS

Mr

0.40

zio

0.38

Mr

0.38

maketitle

0.37

 doctor

0.35

ane

0.35

Master

0.35

Gen

0.35

 Gesch

0.35

Dr

0.34

Activations Density 0.009%

titles and honorifics used to refer to individuals

titles preceding names

No Comments

No Known Activations