INDEX

Explanations

references to potential conditions, operations, or properties related to various subjects

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

possibility or frequency

np_acts-logits-general · gemini-2.0-flash

many of them

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

getDoctrine

-0.36

ряд

-0.34

 Managua

-0.33

 instead

-0.33

 Sache

-0.32

da

-0.32

normal

-0.32

intios

-0.32

 everything

-0.31

instead

-0.31

POSITIVE LOGITS

 للاسماء

0.63

featureID

0.62

ValueStyle

0.61

QMetaType

0.57

 InputDecoration

0.54

 nonatomic

0.54

GEBURTS

0.52

OGND

0.52

GraphicsUnit

0.52

 Signalez

0.52

Activations Density 0.449%

references to potential conditions, operations, or properties related to various subjects

possibility or frequency

many of them

No Comments

No Known Activations