INDEX

Explanations

references to idealized places or states of happiness, often described using terms associated with paradise, utopia, or bliss

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

ideal place or state

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

Bill

-0.45

 Spiders

-0.42

 полиции

-0.42

 Bill

-0.41

William

-0.40

Spider

-0.40

meuble

-0.40

Scooby

-0.40

Shirts

-0.40

Putin

-0.40

POSITIVE LOGITS

 paradise

1.12

 Paradise

1.05

paradise

0.98

Paradise

0.98

 Eden

0.79

 PARAD

0.79

 paradiso

0.78

 paraíso

0.75

 Paradis

0.73

Eden

0.71

Activations Density 0.009%

references to idealized places or states of happiness, often described using terms associated with paradise, utopia, or bliss

ideal place or state

No Comments

No Known Activations