INDEX

Explanations

references to projects or project-related terms

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

project

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



-0.63

XmlAccessorType

-0.57

 uVar

-0.56

 handleMessage

-0.56

+#+#

-0.55

upné

-0.54

MethodManager

-0.50

CDATA

-0.50

GEBURTSDATUM

-0.48

 defStyleAttr

-0.47

POSITIVE LOGITS

 project

1.35

project

1.26

Project

1.25

 Project

1.24

 projects

1.05

getProject

1.05

PROJECT

1.05

ProjectName

1.05

 projekt

1.02

 proyecto

1.00

Activations Density 0.061%

references to projects or project-related terms

project

No Comments

No Known Activations