INDEX

Explanations

proper nouns, specifically names and titles

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

Germain Guislain

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

ConstraintMaker

-0.35

ChildScrollView

-0.33

baomidou

-0.31

jPanel

-0.30

tem

-0.29

mé

-0.28

<thead>

-0.28

 capsule

-0.27

may

-0.26

job

-0.26

POSITIVE LOGITS

Germain

1.31

 Germain

1.12

 germain

0.81

featureID

0.77

NameInMap

0.77

Efq

0.71

ValueStyle

0.71

 AttributeSet

0.70

 otomatig

0.70

enumii

0.67

Activations Density 0.000%

proper nouns, specifically names and titles

Germain Guislain

No Comments

No Known Activations