INDEX

Explanations

instances of authorship or writing attribution within a document

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



-0.59

 Administrativna

-0.57

ftagPool

-0.52

Yeet

-0.52

 ModelExpression

-0.50

󠁣

-0.50

 Prieto

-0.50

eradish

-0.48

 nobly

-0.48

tanleria

-0.48

POSITIVE LOGITS

setVerticalGroup

0.60

 AssemblyCulture

0.58

MemoryWarning

0.50

ResumeLayout

0.48

ⓧ

0.44

[*]

0.43

клопе

0.38

 jScrollPane

0.36

IPH

0.35

 setSelected

0.35

Activations Density 0.001%

instances of authorship or writing attribution within a document

No Comments

No Known Activations