INDEX

Explanations

personal expressions and informal dialogue

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

direct, conversational first-person dialogue—especially informal quotes marked by quotation punctuation and contractions.

oai_token-act-pair · gpt-5 Triggered by @jyhe0408

quotations or direct speech, particularly at the beginning of a quoted statement.

oai_token-act-pair · claude-4-5-sonnet Triggered by @jyhe0408

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

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

 néhány

-0.30

 Gegenstand

-0.28

 recomendamos

-0.28

 quedando

-0.28

 colgantes

-0.27

 Angabe

-0.26

 ebenfalls

-0.26

 plateado

-0.26

 kecil

-0.26

 fakta

-0.25

POSITIVE LOGITS

 defaultstate

0.91

 Infórmanos

0.85

principalColumn

0.83

 للمعارف

0.79

 Chwiliwch

0.76

Personendaten

0.74

transQ

0.74

 мәкал

0.73

 kaarangay

0.73

 незавершена

0.73

Activations Density 0.133%

personal expressions and informal dialogue

direct, conversational first-person dialogue—especially informal quotes marked by quotation punctuation and contractions.

quotations or direct speech, particularly at the beginning of a quoted statement.

No Comments

No Known Activations

personal expressions and informal dialogue

direct, conversational first-person dialogue—especially informal quotes marked by quotation punctuation and contractions.

quotations or direct speech, particularly at the beginning of a quoted statement.

No Comments

No Known Activations