INDEX

Explanations

phrases related to indulgence and self-care

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

indulge in or yourself

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

 blacksmith

-0.50

Rena

-0.49

 Hauts

-0.47

naeus

-0.46

AppComponent

-0.46

Team

-0.44

 fences

-0.44

 Blackman

-0.44

IconModule

-0.43

 Team

-0.42

POSITIVE LOGITS

 indulgence

0.85

indul

0.78

 Indul

0.76

 indulge

0.75

 indulgent

0.73

luxury

0.67

 indulged

0.67

 indulging

0.63

 luxurious

0.63

 Luxury

0.63

Activations Density 0.007%

phrases related to indulgence and self-care

indulge in or yourself

No Comments

No Known Activations