INDEX

Explanations

references to core components or functions in a programming or technical context

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 âgées

-0.22

こともあります

-0.22

 serious

-0.22

ientôt

-0.22

 ayudarte

-0.20

imanapun

-0.20

 vainqueur

-0.20

 turut

-0.19

 âgé

-0.19

 âmes

-0.18

POSITIVE LOGITS

 AssemblyCulture

0.86

/**

0.83

久しぶ

0.83

RenderAtEndOf

0.79

demic

0.78

 パンチラ

0.77

<unused43>

0.77

<unused41>

0.77

<unused23>

0.77

<unused42>

0.76

Activations Density 0.045%

references to core components or functions in a programming or technical context

No Comments

No Known Activations

references to core components or functions in a programming or technical context

No Comments

No Known Activations