INDEX

Explanations

specific types or categories in data structures and measurements related to health and physical attributes

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

lipoprotein, LDL, Sodium, Mg, Width, Sound, bass

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

 autorytatywna

-0.95

featureID

-0.86

 surla

-0.77

ValueStyle

-0.75

Jeografia

-0.71

 AssemblyTitle

-0.68

oneofs

-0.68

NameInMap

-0.68

 Савезне

-0.66

GEBURTSDATUM

-0.66

POSITIVE LOGITS

 اكتوبر

0.31

 hebt

0.30

est

0.28

 uren

0.28

 faut

0.26

 grating

0.25

gik

0.25

APB

0.25

々の

0.25

0.24

Activations Density 0.047%

specific types or categories in data structures and measurements related to health and physical attributes

lipoprotein, LDL, Sodium, Mg, Width, Sound, bass

No Comments

No Known Activations