INDEX

Explanations

references to specific measurable qualities and expectations within a context

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

adoption or redemption

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

XmlElement

-0.40

 Cientí

-0.34

WireFormat

-0.31

รม

-0.31

폭

-0.30

createCanvas

-0.29

UpInside

-0.29

 artist

-0.29

XmlAttribute

-0.29

 Travail

-0.28

POSITIVE LOGITS

 AssemblyCulture

0.68

الحياه

0.65

Personensuche

0.65

OGND

0.64

 صوتيه

0.63

 للمعارف

0.59

 ivelany

0.58

 resourceCulture

0.58

 '\\;'

0.57

parsedMessage

0.53

Activations Density 0.062%

references to specific measurable qualities and expectations within a context

adoption or redemption

No Comments

No Known Activations