INDEX

Explanations

references to titles and cards in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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

 estekak

-0.73

 autorytatywna

-0.61

ScopeManager

-0.56

 feroit

-0.53

 informée

-0.53

ReusableCell

-0.52

getDoctrine

-0.52

 חיצוניים

-0.52

 Савезне

-0.51

 Traducción

-0.51

POSITIVE LOGITS

setTitle

0.46

title

0.42

 title

0.40

centerTitle

0.40

 getTitle

0.39

getTitle

0.35

Personendaten

0.35

 Title

0.34

Title

0.33

 setTitle

0.33

Activations Density 0.011%

references to titles and cards in text

No Comments

No Known Activations