INDEX

Explanations

references to Prince Harry and Meghan Markle

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 33-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_33/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.33.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 restoran

-0.40

findpost

-0.36

 właścic

-0.35

 kanton

-0.35

glClear

-0.35

stateProvider

-0.34

ngdoc

-0.33

 festival

-0.31

 prefeitura

-0.31

 owner

-0.31

POSITIVE LOGITS

 royal

0.81

 royals

0.78

 Prince

0.76

Prince

0.75

royal

0.69

 Royals

0.68

Royal

0.67

 Meghan

0.66

 Windsor

0.65

 Duchess

0.64

Activations Density 0.225%

references to Prince Harry and Meghan Markle

No Comments

No Known Activations