INDEX

Explanations

negative values concerning various conditions or 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

存于互联网档案馆

-0.65

multer

-0.60

 CURIAM

-0.57

LElement

-0.56

 quedado

-0.52

 väli

-0.52

CloseOperation

-0.51

parsedMessage

-0.51

CONSIN

-0.51

uğ

-0.49

POSITIVE LOGITS

LookAnd

0.43

Jîn

0.39

RenderAtEndOf

0.37

 Heim

0.33

 Wett

0.32

 Public

0.32

 vesta

0.32

0.31

Segoe

0.31

 Willow

0.31

Activations Density 0.095%

negative values concerning various conditions or attributes

No Comments

No Known Activations