INDEX

Explanations

references to numerical data or rankings

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

words followed by specific conjunctions or next 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

Бахар

-0.62

 OnTriggerEnter

-0.52

 layui

-0.52

ViewImports

-0.51

AnimationsModule

-0.50

providedIn

-0.47

 '\\;'

-0.46

Населення

-0.44

 namanya

-0.44

waitKey

-0.44

POSITIVE LOGITS

 kaarangay

0.38

Jereo

0.38

 Italijanski

0.37

ngu

0.36

følge

0.35

 наве

0.35

 ivelany

0.35

Хьажоргаш

0.34

tagHelperRunner

0.34

NOPQRST

0.33

Activations Density 0.095%

references to numerical data or rankings

words followed by specific conjunctions or next tokens

No Comments

No Known Activations