INDEX

Explanations

specific named entities, focusing on teams or organizations

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

determiners followed by nouns

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

 Italijani

-0.56

ftagPool

-0.53

탤

-0.52

parsedMessage

-0.50

 ilma

-0.50

styleType

-0.49

Хьажоргаш

-0.49

WriteBarrier

-0.48

⤹

-0.48

Jîn

-0.48

POSITIVE LOGITS

HtmlAttribute

0.40

 viewDidLoad

0.40

Hauptartikel

0.39

 Drapeau

0.38

AutoresizingMask

0.37

 echoed

0.37

BeginInit

0.35

ForResult

0.34

 <<<<<<<<<<<<<<

0.34

ThemeOverlay

0.34

Activations Density 0.011%

specific named entities, focusing on teams or organizations

determiners followed by nouns

No Comments

No Known Activations