INDEX

Explanations

phrases indicating purpose or justification in a discussion

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 12-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_12/width_16k/average_l0_130

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.12.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

 cielos

-0.36

 cuartos

-0.35

 notificaciones

-0.33

 totalidad

-0.32

 pasos

-0.32

 usually

-0.32

 collè

-0.32

 contigo

-0.32

 möglichst

-0.31

 enggak

-0.31

POSITIVE LOGITS

ThroughAttribute

0.81

ChildScrollView

0.81

 Administrativna

0.72

tvguidetime

0.68

uxxxx

0.68

 GenerationType

0.67

 laſſen

0.66

 betweenstory

0.66

ſy

0.66

 ſehr

0.65

Activations Density 0.812%

phrases indicating purpose or justification in a discussion

No Comments

No Known Activations