INDEX

Explanations

concepts related to language structure and expression

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

(&:

-0.42

 Great

-0.42

Great

-0.39

 Plenum

-0.38

 जाती

-0.38

 Osna

-0.37

mian

-0.37

 kicking

-0.37

 Spade

-0.37

 civi

-0.36

POSITIVE LOGITS

 مشين

0.52

الدراسه

0.47

extAlignment

0.46

IndentedString

0.43

yntaxException

0.42

ValueStyle

0.42

LookAnd

0.40

 tárgy

0.40

UnsafeEnabled

0.40

Хьажоргаш

0.40

Activations Density 0.026%

concepts related to language structure and expression

No Comments

No Known Activations