INDEX

Explanations

references to historical figures and events related to leadership and governance

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_33/width_131k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.33.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

ungsver

-0.34

-0.32

bu

-0.31

‘

-0.30

 preso

-0.29

 extré

-0.27

“

-0.27

 presa

-0.27

 facing

-0.26

 drink

-0.26

POSITIVE LOGITS

 disambiguazione

0.87

LookAnd

0.74

 initComponents

0.73

 doInBackground

0.70

SequentialGroup

0.70

ArgsConstructor

0.68

WireFormatLite

0.66

RegressionTest

0.65

<unused14>

0.64

<unused51>

0.64

Activations Density 0.022%

references to historical figures and events related to leadership and governance

No Comments

No Known Activations