INDEX

Explanations

phrases related to entering or submitting information

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

verifyException

-0.73

featureID

-0.71

MLLoader

-0.71

ItemBackground

-0.66

sizeCache

-0.63

tagHelperRunner

-0.63

AnchorStyles

-0.62

 tartalomajánló

-0.61

tableFuture

-0.60

 ModelRenderer

-0.60

POSITIVE LOGITS

 entry

0.63

 entered

0.62

Entering

0.60

 Entry

0.59

 Entering

0.56

nEnter

0.56

entered

0.55

 Entered

0.55

 input

0.54

Entry

0.54

Activations Density 0.032%

phrases related to entering or submitting information

No Comments

No Known Activations