INDEX

Explanations

references to news outlets and media organizations

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

news sources

np_acts-logits-general · gemini-2.0-flash

news publications

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

vician

-0.50

 autorytatywna

-0.50

abestanden

-0.39

delijk

-0.38

 consultations

-0.38

 çalışmalar

-0.38

rhosis

-0.38

ruptedException

-0.38

 RADIATION

-0.37

탤

-0.37

POSITIVE LOGITS

ьаж

0.45

 ligiloj

0.44

 beginnetje

0.43



0.43

RegistryLite

0.43

 magazine

0.40

/***/

0.40

 tabloid

0.40

Zir

0.40

脚注の使い方

0.40

Activations Density 0.136%

references to news outlets and media organizations

news sources

news publications

No Comments

No Known Activations