INDEX

Explanations

references to programming language packages or modules

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

java code and library paths

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

 Administración

-0.41

Personendaten

-0.40

 död

-0.40

 Infór

-0.40

 événement

-0.40

 diable

-0.40

 dumne

-0.40

 opinión

-0.39

 dead

-0.38

 Realität

-0.38

POSITIVE LOGITS

IUrlHelper

0.72

InputBorder

0.63

StoryboardSegue

0.57

fromnode

0.56

 AssemblyCompany

0.55

 ProtoMessage

0.54

 QtWidgets

0.54

TestingModule

0.53

IBOutlet

0.50

LookAnd

0.49

Activations Density 0.001%

references to programming language packages or modules

java code and library paths

No Comments

No Known Activations