INDEX

Explanations

instances of opening or starting phrases, or segments in text

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

the beginning of new sentences or segments

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

numbers and codes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

 kasarigan

-0.47

 recently

-0.46

 kürzlich

-0.46

 Anfang

-0.46

 anfangs

-0.46

ITOS

-0.45

 récents

-0.45

 autrefois

-0.44

 earlier

-0.44

 noong

-0.43

POSITIVE LOGITS

PhysRevD

0.72

 bezeichneter

0.71

 مرئيه

0.70

OGND

0.66

setupUi

0.65

λίου

0.63

styleable

0.62

ingeki

0.62

preduce

0.60

 виправивши

0.59

Activations Density 0.052%

instances of opening or starting phrases, or segments in text

the beginning of new sentences or segments

numbers and codes

No Comments

No Known Activations