INDEX

Explanations

topics related to health, science, and technology

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

modifier and noun concept

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

 Wohnungen

-0.57

 poszczegól

-0.52

 Entscheidungen

-0.52

 Öffnungszeiten

-0.51

 egyes

-0.51

 Verhandlungen

-0.50

 dignité

-0.50

 Versuche

-0.50

 apropiado

-0.50

 Übersetzungen

-0.49

POSITIVE LOGITS

elemField

0.68

 estekak

0.64

 zuſammen

0.62

GEBURTSDATUM

0.60

 tartalomajánló

0.60

 basket

0.60

>\<^

0.59

 istr

0.59

 regime

0.58

цездатний

0.58

Activations Density 0.163%

topics related to health, science, and technology

modifier and noun concept

No Comments

No Known Activations