INDEX

Explanations

instances of the word "new," indicating a focus on new object creation in programming contexts

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

instantiating classes

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

 الحره

-0.77

 betweenstory

-0.69

tagHelperRunner

-0.68

TagHelper

-0.58

transQ

-0.47

 indígen

-0.47

 незавершена

-0.46

-------------</

-0.45

AnchorTagHelper

-0.45

 eccell

-0.44

POSITIVE LOGITS

WireFormatLite

0.57

wapV

0.51

new

0.50

new

0.48

 تضيفلها

0.46

SOUNDBITE

0.45

migo

0.43

RouterModule

0.42

ecore

0.41

ANSA

0.40

Activations Density 0.002%

instances of the word "new," indicating a focus on new object creation in programming contexts

instantiating classes

No Comments

No Known Activations

instances of the word "new," indicating a focus on new object creation in programming contexts

instantiating classes

No Comments

No Known Activations