INDEX

Explanations

direct speech or quotes in the text

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

dialogue or quotations in the text

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

reporting jokes or laughter

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

IonicModule

-0.78

MigrationBuilder

-0.76

 ویکی‌پدیا

-0.74

bootstrapcdn

-0.67

HomeAsUpEnabled

-0.66

ագրություններ

-0.65

 للمعارف

-0.65

اقرأ

-0.64

 oprot

-0.64

っていきます

-0.64

POSITIVE LOGITS

 laughs

1.67

 laugh

1.65

 joked

1.62

 laughed

1.61

 laughing

1.50

 joke

1.43

 jokingly

1.43

 jokes

1.41

 joking

1.40

 laughter

1.37

Activations Density 0.019%

direct speech or quotes in the text

dialogue or quotations in the text

reporting jokes or laughter

No Comments

No Known Activations