INDEX

Explanations

phrases indicating sources or authorship in a text

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

prepositions

np_max-act · gemini-2.0-flash Triggered by @johnny

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_1m/average_l0_107

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

1,048,576

Data Type

float32

Hook Name

blocks.12.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

EndProject

-0.60

 Schenk

-0.56

StructEnd

-0.54

 disambiguazione

-0.53

ứa

-0.52

Eksteraj

-0.52

homonymie

-0.50

nitten

-0.49

kesha

-0.48

 geldig

-0.48

POSITIVE LOGITS

bootstrapcdn

0.76



0.65

 DialogInterface

0.60

 kautta

0.59

endphp

0.58

 AttributeSet

0.57

offie

0.54

 فريبيس

0.53

seded

0.52

 edili

0.52

Activations Density 0.001%

phrases indicating sources or authorship in a text

prepositions

No Comments

No Known Activations