INDEX

Explanations

elements associated with programming logic and structure

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

`min`, `Pub`, `is`, `INST`, `max`, `run`, `local` followed by specific tokens

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

fortawesome

-0.44

 IBOutlet

-0.43

hyrchwyd

-0.42

insatz

-0.40

 становника

-0.38

存于互联网档案馆

-0.37

cèse

-0.37

GMENT

-0.36

forKey

-0.36

 Thirdly

-0.36

POSITIVE LOGITS

Aiheesta

0.45

囗

0.44

AppMethodBeat

0.39

tagext

0.38

inig

0.38

Спасылкі

0.37

 BrowserModule

0.36

Примечания

0.36

叶修

0.36

Clik

0.36

Activations Density 0.144%

elements associated with programming logic and structure

`min`, `Pub`, `is`, `INST`, `max`, `run`, `local` followed by specific tokens

No Comments

No Known Activations