INDEX

Explanations

humorous elements and self-deprecating remarks in discussions

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

Humor, parody, or sarcasm

np_acts-logits-general · gemini-2.0-flash

joking, humor, and sarcasm

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

humorous, tongue-in-cheek content and references to comedy, including jokes, quips, puns, parody, and self-deprecating roasts.

oai_token-act-pair · gpt-5 Triggered by @vetterc0

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

ilíbrio

-0.32

 conforto

-0.31

घ

-0.30

 Catawiki

-0.29

 StatefulWidget

-0.29

Distribuzione

-0.29

ampiyon

-0.29

ঘ

-0.29

XmlAccessorType

-0.28

 Rain

-0.28

POSITIVE LOGITS

 parody

0.73

 sarcas

0.67

 joking

0.66

 jokingly

0.65

 humor

0.65

 satirical

0.64

 joked

0.63

 humour

0.63

 humorous

0.63

 jokes

0.62

Activations Density 0.204%

humorous elements and self-deprecating remarks in discussions

Humor, parody, or sarcasm

joking, humor, and sarcasm

humorous, tongue-in-cheek content and references to comedy, including jokes, quips, puns, parody, and self-deprecating roasts.

No Comments

No Known Activations