INDEX

Explanations

phrases indicating location or arrangement

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-mlp/layer_6/width_131k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.6.hook_mlp_out

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

 aporta

-0.33

 kual

-0.32

 éste

-0.32

 skla

-0.31

 škoda

-0.31

 Erfahrungen

-0.31

 Behaup

-0.30

 commitments

-0.30

TableHead

-0.30

 impresiones

-0.30

POSITIVE LOGITS

 autorytatywna

1.11

rungsseite

0.81

OGND

0.80

Diweddarwch

0.76

hoeddwyd

0.75

oa̍t

0.74

RegressionTest

0.73

 disambiguazione

0.71

 Италијани

0.70

 poffible

0.69

Activations Density 0.102%

phrases indicating location or arrangement

No Comments

No Known Activations