INDEX

Explanations

phrases indicating the state of being or existence

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

forms of "to be" followed by description

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

 éton

-0.57

 forfait

-0.56

spyOn

-0.51

horesis

-0.50

 şöyle

-0.50

 descrizione

-0.49

 tanong

-0.49

 gyhoeddwyd

-0.49

 segno

-0.48

 aján

-0.47

POSITIVE LOGITS

 becoming

0.46

лтемелер

0.46

not

0.43

enumerate

0.40

 бъде

0.39

]]>

0.39

 doing

0.39

 legitimate

0.39

 Administrativna

0.38

OGND

0.38

Activations Density 0.096%

phrases indicating the state of being or existence

forms of "to be" followed by description

No Comments

No Known Activations