INDEX

Explanations

elements related to code or programming constructs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 31-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_31/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.31.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

featureID

-0.79

 الحره

-0.78

ﬄ

-0.77

oa̍t

-0.74

Jeografia

-0.73

Tembelea

-0.71

adaptiveStyles

-0.70

 Administrativna

-0.67

ImageContext

-0.67

 كومونز

-0.66

POSITIVE LOGITS

 extranjera

0.51

 izquierdo

0.51

 cooperación

0.51

 duque

0.50

 Italij

0.48

 Económica

0.47

 restauración

0.46

 desierto

0.45

 Relaciones

0.45

 visitante

0.43

Activations Density 0.064%

elements related to code or programming constructs

No Comments

No Known Activations