INDEX

Explanations

key names and associated contact or relational information

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

Between parentheses or after a semicolon

np_acts-logits-general · gemini-2.0-flash

structured contact or description following specific keywords

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 myſelf

-0.64

 purpoſe

-0.59

 houſe

-0.57

 cauſe

-0.57

inafter

-0.56

 --------------

-0.56

 preſent

-0.54

 perſon

-0.54

 inſ

-0.53

 δὲ

-0.52

POSITIVE LOGITS

www

0.74

www

0.63

onedDateTime

0.60

Personendaten

0.56

 المعيارى

0.56

 http

0.54

 StatelessWidget

0.53

ENE

0.53

lapsingToolbar

0.52

 réjou

0.52

Activations Density 0.178%

key names and associated contact or relational information

Between parentheses or after a semicolon

structured contact or description following specific keywords

No Comments

No Known Activations