INDEX

Explanations

negative conjugated forms of the verb "to be."

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-mlp/layer_6/width_131k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.6.hook_mlp_out

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

 wont

-1.11

 wouldnt

-0.71

 arent

-0.67

 couldnt

-0.63

 isnt

-0.60

 wasnt

-0.59

 doesnt

-0.57

 didnt

-0.55

Won

-0.54

 cant

-0.52

POSITIVE LOGITS

rungsseite

0.63

RegressionTest

0.50

+#+

0.48

AddHtmlAttribute

0.47

0.46

stylers

0.46

NameInMap

0.46

shutil

0.43

//

0.41

createCanvas

0.41

Activations Density 0.001%

negative conjugated forms of the verb "to be."

No Comments

No Known Activations