INDEX

Explanations

references to royal or powerful figures and their attributes

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

CrossRef

-0.49

 Brod

-0.36

른

-0.34

Sop

-0.33

 suficientes

-0.33

 Söhne

-0.33

)}.

-0.33

Pun

-0.32

 Home

-0.32

 amenable

-0.32

POSITIVE LOGITS

MemoryWarning

0.67

новништво

0.61

 كومونز

0.56

hoeddwyd

0.54

SharedCtor

0.51

AntiForgeryToken

0.50

 الرياضيه

0.50

 Normdatei

0.48

verwijspagina

0.48

تقاوى

0.48

Activations Density 0.009%

references to royal or powerful figures and their attributes

No Comments

No Known Activations