INDEX

Explanations

phrases indicating ownership or personalization

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Tatsache

-0.51

 enfans

-0.48

 faptul

-0.48

jména

-0.46

 wzglę

-0.45

Referencer

-0.43

 naciones

-0.42

PostBack

-0.42

 salariés

-0.42

 השאלה

-0.42

POSITIVE LOGITS

own

0.59

ftagPool

0.57

GHIJKLM

0.56

 мәкал

0.56

atoday

0.53

 jigsaw

0.52

存于互联网档案馆

0.52

DIY

0.50

Own

0.50

 smarter

0.50

Activations Density 0.010%

phrases indicating ownership or personalization

No Comments

No Known Activations