INDEX

Explanations

positive adjectives and evaluative phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 10-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_10/width_16k/average_l0_113

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.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

цездатний

-0.92

 betweenstory

-0.82

MigrationBuilder

-0.69

IVEREF

-0.67

NameInMap

-0.66

richTextPanel

-0.66

 estekak

-0.66

rungsseite

-0.65

RTGC

-0.64

########.

-0.62

POSITIVE LOGITS

 schwed

0.33

 sağ

0.33

 bağlı

0.33

 geçir

0.31

":"",

0.30

SuppressLint

0.29

 znaj

0.29

"]/

0.29

 sağlay

0.28

 guerres

0.28

Activations Density 1.568%

positive adjectives and evaluative phrases

No Comments

No Known Activations