INDEX

Explanations

references to global concepts and phenomena

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

"global"

np_acts-logits-general · gemini-2.0-flash

global contexts

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

LookAnd

-0.46

 defStyleAttr

-0.45

œuvre

-0.45

 getRule

-0.44

 MenuView

-0.43

étoit

-0.43

Medicine

-0.43

 ViewController

-0.43

 setDescription

-0.42

ErrorException

-0.42

POSITIVE LOGITS

 global

1.15

Global

1.07

global

1.06

 Global

1.05

 GLOBAL

1.04

GLOBAL

0.99

lobal

0.91

 globale

0.82

 глоба

0.76

全局

0.74

Activations Density 0.048%

references to global concepts and phenomena

"global"

global contexts

No Comments

No Known Activations