INDEX

Explanations

terms related to artificiality and deception

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

"Fake" or artificial things

np_acts-logits-general · gemini-2.0-flash

artificial or fake followed by nouns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

tamment

-0.63

 pribadi

-0.51

usercontent

-0.50

ತ್ತು

-0.49

{\

-0.49

masing

-0.49

국의

-0.48

Населення

-0.47

ಂತ

-0.47

 lourd

-0.46

POSITIVE LOGITS

 Fake

0.72

 sembl

0.66

 pretence

0.66

 pretense

0.66

denza

0.64

 fake

0.64

Fake

0.63

 Mock

0.62

Faux

0.62

 pretend

0.62

Activations Density 0.199%

terms related to artificiality and deception

"Fake" or artificial things

artificial or fake followed by nouns

No Comments

No Known Activations