INDEX

Explanations

proper nouns, particularly names of individuals and places

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

names

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

rzez

-0.54

EFIAPI

-0.48

 Bewußt

-0.44

écnicas

-0.44

 daß

-0.43

getahuan

-0.43

фициальный

-0.42

wußt

-0.41

 flesta

-0.41

 lesbians

-0.41

POSITIVE LOGITS

ArgsConstructor

0.66

Datuak

0.56

KURZBESCHREIBUNG

0.54

 betweenstory

0.53

}));

0.51

 Astor

0.50

AndroidJUnit

0.50

XmlAccessorType

0.49

ModelSerializer

0.48

 HasFactory

0.48

Activations Density 0.080%

proper nouns, particularly names of individuals and places

names

No Comments

No Known Activations